Agenda packet — Ava Community Energy Board of Directors, February 18, 2026

Board of Directors Meeting
Wednesday, February 18, 2026
6:00 pm
In Person:

Board Room
Ava Community Energy
1999 Harrison St, Ste 2300
Oakland, CA 94612
Or from the following locations:
• Clipper Club - 5 Captain Dr. Emeryville, CA 94608
• 4917 Knowlson Terrace, Fremont, CA 94555
• 35653 Scarborough Dr Newark, CA 94560
• 1100 Webster St, 2nd Floor Conference Room, Oakland, CA 94607 • Larch Clover Community Center – 11157 West Larch Road, Tracy, CA 95376 • Stockton City Hall – 425 N El Dorado St., Stockton, CA 95202 Via Zoom:
https://avaenergy-org.zoom.us/j/87023071843
Or join by phone:

Dial(for higher quality, dial a number based on your current location): US: +1 669 900 6833 or +1 346 248 7799 or +1 253 215 8782 or +1 929 205 6099 or +1 301 715 8592 or 888 475 4499 (Toll Free) or 877 853 5257 (Toll Free)
Webinar ID: 870 2307 1843
Meetings are accessible to people with disabilities. Individuals who need special assistance or a disability-related modification or accommodation to participate in this meeting, or who have a disability and wish to request an alternative format for the meeting materials, should contact the Clerk of the Board at least 2 working days before the meeting at (510) 906- 0491 or [email protected].

If you have anything that you wish to be distributed to the Board of Directors, please email it to the clerk by 5:00 pm the day prior to the meeting.

1. Welcome & Roll Call
Ava Community Energy
Board of Directors Page | 2

2. Pledge of Allegiance

3. Public Comment
This item is reserved for persons wishing to address the Board on any Ava-related matters that are not otherwise on this meeting agenda. Public comments on matters listed on the agenda shall be heard at the time the matter is called. As with all public comment, members of the public who wish to address the Board are customarily limited to two minutes per speaker and must complete an electronic speaker slip. The Board Chair may increase or decrease the time allotted to each speaker.

CONSENT AGENDA

4. Approval of Minutes from January 21, 2026

5. Contracts Entered Into (Informational Item)

6. FY 2025-2026 Q2 Treasurer’s Report
Update on Ava’s for FY 2026 Q2 cash position

7. Contract Amendment to ESCA-PLD-Tracy, 9 LLC
Contract amendment to the DAC project ESCA-PLD-TRACY9, LLC

8. Contract Amendment to ESCA-PLD-Tracy, 16 LLC
Contract amendment to the DAC project ESCA-PLD-TRACY16, LLC

REGULAR AGENDA

9. CEO Report

10.CAC Report

11.Fiscal Year 2025-2026 Mid-Year Budget Review (Informational Item) Informational item presenting the mid-year performance of the current FY budget

12.Overview of large electric load growth trends and implications for Ava Community Energy (Informational Item)
Informational Item providing an overview of large electric load growth trends and implications for Ava Community Energy

13.SmartHome Battery Program Launch Update (Informational Item) Informational items presenting program development and status before program launch.

14.Board Member and Staff Announcements including requests to place items on future Board of Directors Meeting Agendas
Ava Community Energy
Board of Directors Page | 3

15.Adjourn
The next Ava Board of Directors meeting will be held on Wednesday, March 18, 2026 at 6pm.

Board Meeting Access Instructions
If you need help finding or accessing the building, please call our Ava representative who is stationed in the building lobby: 510-393-0492.

Directions
Directions via BART
If you are taking BART: the 19th Street station is the closest stop to our office and is about a 5 minute walk away. Use the 20th St / Thomas L. Berkeley Way station exit.

Directions via Bike
Bike riders wanting to park their bike inside the parking garage can enter through the main building lobby. Bike parking is available on the parking garage first level right in front of the garage elevators.

Directions to Parking Garage via Car
The entrance to the building’s attached garage is located on Harrison Street. If you’re driving northbound on Harrison Street, as soon as you cross 19th St. the garage entrance is 3/4 down on your left-hand side. If you’re heading east on Thomas L Berkeley Way/20th St.

Continue East then make a right turn on Harrison
Street, and the garage entrance is a quarter block up
on your right-hand side.

When you arrive, enter via the gate labeled “Public
Parkers”. There are four floors of the parking garage,
and you will need to take the elevator in the parking
garage to the first floor. The parking attendant or
an Ava representative will provide access into the
building lobby.

The parking garage entry gate will be open until 8pm
for CAC and BOD meetings. Attendees can exit the
parking garage until 11pm.

Note that the garage’s parking fee is $30 per use.

Street parking is widely available near the building
and free after 6pm.

Check-in at Security Desk
When you arrive at our building, please check in with the security desk in the lobby to get access to the elevators.

If you have questions or need assistance, an Ava representative will be stationed and identifiable in the lobby. They can be reached at: 510-393-0492.

Public Parkers entrance
1999 Harrison Street, Suite 2300, Oakland, CA 94612

Draft Minutes
Board of Directors Meeting
Wednesday, January 21, 2026
6:00 pm
In Person:

Board Room
Ava Community Energy
1999 Harrison St, Ste 2300
Oakland, CA 94612
Or from the following locations:
• Office of Councilmember Igor Tregub, 2180 Milvia St., 5th Floor, Berkeley, CA 94704
• Hyatt Regency Sacramento, 1209 L Street, Sacramento, CA 95814 • 4917 Knowlson Terrace, Fremont, CA 94555
• 35653 Scarborough Dr Newark, CA 94560
• 1100 Webster St, 2nd Floor Conference Room, Oakland, CA 94607 • Larch Clover Community Center – 11157 West Larch Road, Tracy, CA 95376 • Stockton City Hall – 425 N El Dorado St., Stockton, CA 95202 • 1755 Harvest Landing Ln., Tracy, CA 95376
Via Zoom:
https://avaenergy-org.zoom.us/j/87023071843
Or join by phone:

Ava Community Energy
Board of Directors Page | 2
If you have anything that you wish to be distributed to the Board of Directors, please email it to the clerk by 5:00 pm the day prior to the meeting.

1. Welcome & Roll Call

Present: Members: Marquez (Alameda County), Tregub (Berkeley), Kaur (Emeryville), Keng (Fremont), Roche (Hayward), Barrientos (Livermore), Del Catancio (Newark), Brown (Oakland), Balch (Pleasanton), Rickman (San Joaquin County), Fugazi (Stockton), Bedolla (Tracy), Sakakihara (Union City), CAC Chair Souza (Community Advisory Committee), Vice-Chair Gonzalez (San Leandro) and Chair Andersen (Piedmont) Not Present: Members: Lopez (Albany), Morada (Dublin) and Diallo (Lathrop) Alternate Member Bedolla served as the representative for the City of Tracy (Nygard).

Member Kaur joined the meeting at 6:30pm.

Alternate Member Bedolla joined the meeting at 6:43pm.

Member Barrientos joined the meeting at 7:11pm.

2. (2:05) Pledge of Allegiance
The Chair led the Pledge of Allegiance.

CONSENT AGENDA

4. (5:41) Approval of Minutes from December 17, 2025

5. Contracts Entered Into (Informational Item)
Vice-Chair Gonzales motioned to approve the Consent Calendar. Member Roche seconded the motion, which was approved 13/0/0/0/5 (yes/no/abstain/recuse/not present)
Yes: Members: Marquez, Tregub, Kaur, Keng, Roche, Del Catancio, Brown, Balch, Rickman, Fugazi, Sakakihara, Vice-Chair Gonzalez and Chair Andersen No: none
Abstain: none
Recuse: none
Not Present: Members: Lopez, Morada, Diallo, Barrientos, Bedolla Ava Community Energy
Board of Directors Page | 3

REGULAR AGENDA

6. (7:49) CEO Report
Howard Chang, CEO, presented the CEO Report. The written version of the report is available in the January 21, 2026 agenda packet.

Chair Andersen opened the public comment period and there were no speakers.

7. (13:04) CAC Report
The newly elected CAC Chair, Lorraine Sousa, delivered her first report to the Board.

The written version of the report is available in the January 21, 2026 agenda packet.

Chair Andersen opened the public comment period and there were no speakers.

8. (18:42) IRP Overview (Informational Item)
Overview of upcoming Integrated Resource Plan compliance analysis & Ava supplemental analysis
Marie Fontenot, Senior Vice President of Power Resources, presented an overview of the upcoming Integrated Resource Plan (IRP) compliance analysis and answered questions from the Board.
(54:47) Public Comment: Jim Lutz expressed the CAC’s appreciation for staff bringing in house analysis such as that provided in the IRP Overview.

9. (1:02:17) Update on Changes to 2026 Customer Bills (Informational Item) Update on rates and a new PG&E delivery related Base Service Charge Annie Henderson, Chief Customer Officer, presented an update on changes to 2026 customer bills and answered questions from the Board.
(1:33:07) Public Comment: Jessica Tovar stated that in addition to conducting outreach about rates and affordability, there should also be discussion about the benefits that a public not for profit community choice energy agency can bring for local communities.

10.(1:45:25) Meeting Agenda Schedule Protocol (Action Item) Discuss a schedule management protocol to require agenda items to start prior to 9PM for regular board meetings with an option to extend with majority board vote Chair Andersen presented the Meeting Agenda Schedule Protocol and answered questions from the Board.

Ava Community Energy
Board of Directors Page | 4
Member Roche motioned to approve the Meeting Agenda Schedule Protocol.

Member Barrientos seconded the motion, which was approved 14/0/0/0/4 (yes/no/abstain/recuse/not present):

Yes: Members: Marquez, Tregub, Kaur, Keng, Roche, Barrientos, Del Catancio, Brown, Balch, Rickman, Fugazi, Sakakihara, Vice-Chair Gonzalez and Chair Andersen
No: none
Abstain: none
Recuse: none
Not Present: Members: Lopez, Morada, Diallo and Bedolla

11.(1:51:08) Board Member and Staff Announcements including requests to place items on future Board of Directors Meeting Agendas
There were no announcements or requests.

12.Adjourn
The meeting was adjourned at 7:53pm.

The next Ava Board of Directors meeting will be held on Wednesday, February 18, 2026 at 6pm.
1/21/26 – Ava Board of Directors
AI Generated Courtesy Summary
not official minutes
This summary has not been reviewed for accuracy
Welcome & Roll Call
The Board of Directors meeting was called to order at 6:02 PM on Wednesday, January 21, 2026, by the Chair.

The meeting was held at Ava headquarters at 1999 Harrison Street, Suite 2300 in Oakland, California in the Altamonte conference room. The Clerk conducted roll call and established that a quorum was present with representatives from Alameda County, Albany, Berkeley, Dublin, Emeryville, Fremont, Hayward, Lathrop, Livermore, Newark, Oakland, Pleasanton, San Joaquin County, Stockton, Tracy, Union City, Community Advisory Committee, San Leandro, and Piedmont in attendance.

Pledge of Allegiance
The Chair led the Pledge of Allegiance.

Public Comment
Jessica Tovar from Local Clean Energy Alliance and People's Clean Power Alliance addressed the Board. She reminded the Board about the devastating wildfires that occurred a year ago, emphasizing how these events demonstrated the need for community resilience in the face of climate change. She advocated for resilience hubs to respond to current crises and thanked the Board for not moving forward with contracting Deloitte, which holds contracts with ICE and the Department of Homeland Security. She urged the agency to create a policy that prevents public funds from being used to contract with organizations collaborating with what she termed a "fascist regime."

CONSENT AGENDA
Approval of Minutes from December 17, 2025
Contracts Entered Into (Informational Item)
The consent agenda included two items: approval of minutes from December 17, 2025, and contracts entered into by Ava in January 2026.

Motion to approve the consent agenda was made by Vice Chair and seconded by Member Roche. The motion was approved unanimously by roll call vote.

CEO Report
The CEO presented several updates:
● Recent meetings included an Executive Committee meeting on January 14th where members received an update on customer care and data management services procurement plans. The next Executive Committee meeting on February 4th would likely be canceled due to staff being in Sacramento. ● The Financial Administrative and Procurement (FAP) meeting is scheduled for January 28th at 3:30 PM, while the Marketing Regulatory Legislative subcommittee meeting on December 19th was canceled, with the next one scheduled for March 6th.
● Applications for the Community Advisory Committee (CAC) are now open through February 20th.

The CEO welcomed and congratulated Mickey Sousa as the newly elected CAC Chair. ● The CEO mentioned ongoing work on Ava's outreach strategy, with plans to survey board and CAC members for input on 2026 outreach plans.
● Recent community sponsorship recipients were highlighted, with funding provided to organizations across Ava's service area in areas such as sports, art, music, and STEM. ● A grants ideation public community workshop will be held on January 28th from 5:30-7:30 PM at the Oakland office, with a remote option available.

CAC Report
The newly elected CAC Chair, Mickey Sousa, delivered his first report to the Board. He thanked the committee for electing him as Chair alongside Vice Chair Indera Bachlissen. Souza outlined his goals for the year, which included better understanding how the advisory committee can serve both Ava staff and the Board while acting as a bridge to community needs.

The Chair reported that during their recent meeting, Jessica Tovar had emphasized that communities are already suffering from wildfires and climate change issues, and thanked Ava for opposing corporations working with ICE. The CAC was pleased with the addition of AI-generated minutes despite some name misspellings.

The CAC expressed disappointment regarding the reduction of renewable investments and removal of unbilled credits. The committee also discussed streamlining ad hoc committee communications using categories such as risks, community concerns, legislative issues, and opportunities.

Souza highlighted the possibility of joining forces with San Diego CCA, which has sued for a review of the PCIA (Power Charge Indifference Adjustment), suggesting that Ava and the CAC might help with efforts to reverse recent PCIA reform.

IRP Overview (Informational Item)
Marie Fontenot, Senior Vice President of Power Resources, presented an overview of the upcoming Integrated Resource Plan (IRP) compliance analysis. She explained that the IRP is a biennial compliance filing required by the California Public Utilities Commission (CPUC) that serves as a long-term road map detailing how Ava will meet future energy demand reliably and affordably.

Key points from her presentation included:
● The CPUC mandates prescriptive modeling assumptions for all load-serving entities, with the filing deadline set for June 1, 2026.
● Previous IRP analyses in 2018, 2020, and 2022 have resulted in procurement mandates from the CPUC.
● Ava is now using an open-source software called GenX to conduct this analysis in-house for the first time, which represents a significant advancement in organizational capabilities. ● The analysis includes three main modeling components: capacity expansion, production cost, and portfolio expansion.
● The CPUC will require analysis of different carbon emission scenarios, including assumptions about statewide carbon goals by 2040.
● The results will be presented to the Board for approval before submission to the CPUC in May or June.

Howard Choy, CEO, emphasized the significant undertaking this analysis represents and noted that developing these capabilities in-house will allow Ava to perform similar analyses more regularly to inform decision- making. The Board discussed the potential for an ad hoc committee to provide additional review of the IRP materials before they come back for final approval.

Members expressed concerns about forecasting data center loads, which are projected to grow significantly (up to 20% year-over-year from 2027-2029), and questions about whether small-scale nuclear power would be part of the state's energy future.

Update on Changes to Customer Bills (Informational Item) Annie Henderson, Chief Customer Officer, presented an update on changes to 2026 customer bills. She explained that due to a late December filing with final rates for January 2026, there were changes from the forecast presented in December 2025:
● Previously, the total bill decrease for average residential customers was forecasted to be about 11%, but based on final numbers, it will be about 5%.
● Ava is still reducing its generation charges by 15-21% depending on service level, but PG&E's delivery charges have increased more than expected.
● The Bright Choice discount and Renewable 100 premium changes approved by the Board in December have been implemented, with most customers seeing these new rates on bills received in early February.
● A new base services charge (fixed monthly fee) from PG&E will begin appearing on residential customers' bills in March 2026, with standard residential customers paying $24 per month, CARE customers paying $6, and FERA customers and those in subsidized housing paying $12.

Several Board members expressed concern about customer confusion regarding bills and the need for clear communication distinguishing between Ava charges and PG&E charges. Member Fugazi shared that her office receives daily questions about energy bills and requested consistent messaging and clear side-by-side comparisons to help constituents understand their bills.

Vice Chair emphasized the importance of clearly communicating that Ava is working to lower customer bills through its portion of the charges, even if the overall bill has increased due to PG&E charges. Member Tregub noted concerns about the base services charge's potential impacts on lower-income customers, especially those in smaller housing units.

Meeting Agenda Schedule Protocol (Action Item)
The Chair presented a proposed protocol to require agenda items to start prior to 9 PM for regular board meetings, with an option to extend with a majority board vote. The proposal was developed to improve meeting efficiency and help staff prepare accordingly, now that meetings are held at Ava offices without the previous venue's 10 PM cutoff.

Board members were supportive of the proposal, with Member Roche noting that it would encourage items to be completed on time and Member Barantas adding that it would prevent items from being drawn out unnecessarily. The Chair clarified that if a substantive agenda item had not begun by 9 PM, it would be continued to the next meeting unless a majority of members voted to proceed.

Motion to approve the meeting agenda schedule protocol was made by Member Roche and seconded by Member Barantas. The motion was approved unanimously by roll call vote.

Board Member and Staff Announcements including requests to place items on future Board of Directors Meeting Agendas There were no announcements or requests.

Adjourn
The meeting was adjourned at 7:53 PM. The next Ava Board of Directors meeting will be held on Wednesday, February 18, 2026, at 6 PM.

Consent Item 5
Consent Item 5
To: Ava Community Energy Authority
From: Howard Chang, Chief Executive Officer
Subject: Contracts Entered Into
Date: February 18, 2026
_____________________________________________________________________

RECOMMENDATION
Accept the CEO’s report on contracts that Ava Community Energy has entered into, as required by the Administrative Procurement Policy, from January 13th, 2026 through February 13, 2026.

Consent Item 5
Counterparty
Name
Record Name Record
Type
Contract Purpose Contract
Value
Not To
Exceed
Expiration
Date
City State Ironclad
Id
Abbott,
Stringham &
Lynch
January 2026
Consulting
Services
Agreement
Consulting
Services
Agreement
Provides data and
security audit
services to Ava.
$10,000 $10,000 2026-06-
30 Campbell California IC-2278
ACCO
Engineered
Systems
BID CONTRACT Vendor
Purchase
Template
Provide exigent
evaluation of
potential repairs to
the HVAC system in
Ava's server room.
$1,008 N/A N/A San
Leando
CA IC-2296

COASTAL

CLOUD, LLC
January 2026
Amendment to
the Consulting
Services
Agreement
Consulting
Services
Agreement
addition of Scope
related to Marketing
Cloud Growth
Implementation,
adds $63,450 to the
compensation.
$63,450 $213,350 N/A Palm
Coast
Florida IC-2283
Cornish &
Carey
Commercial
February 2026
Amendment to
the California
Exclusive
Authorization
of Sale
Vendor
Purchase
Template
Extend the term of a
real estate broker
agreement until
March 31, 2026.

N/A N/A 2026-03-
31 Oakland California IC-2347
LHi Group,
Inc.

Statement of
Work -
Payrolling
Assignment
Vendor
Purchase
Template
Temporary
workforce
supplementation.

N/A
(Hourly)
N/A 2026-
0630
New York New
York
IC-2302
Magellan
Health
Services of
California, Inc
Services
Agreement
Vendor
Purchase
Template
Employee
Assistance Program.
$30,000 N/A 2029-01-
31 Maryland
Heights
Missouri IC-2284
Consent Item 5
Counterparty
Name
Record Name Record
Type
Contract Purpose Contract
Value
Not To
Exceed
Expiration
Date
City State Ironclad
Id
Noceti Group,
Inc.
2026 San
Joaquin County
Asparagus
Festival $5,000
Spear
Sponsorship
Vendor
Purchase
Template
Sponsorship
provided to support
a local festival in
Ava's service
territory in
exchange for
promotional
privileges.
$5,000 $5,000 N/A French
Camp
California IC-2332
OpenAI OpCo,
LLC

OPENAI ORDER
FORM
Vendor
Purchase
Template
Procure AI licenses
with appropriate
legal protections
and use limitations.
$19,800 N/A 2027-02-
15 San
Francisco
California IC-2457
Port City
Marketing
Solutions,
Incorporated
December
2025
Consulting
Services
Agreement
Consulting
Services
Agreement
Provides media
buying and related
marketing and
communications
services to Ava.
$100,000 $100,000 2027-01-
20 Stockton California IC-2261
Project6
Design, Inc.

February 2026
Amendment to
the Consulting
Services
Agreement
Consulting
Services
Agreement
Increase NTE on
existing contract.
$20,000 $468,000 N/A Berkeley California IC-2345

Consent Item 6
Consent Item 6
To: Ava Community Energy Authority
From: Russell Mills, Chief Financial Officer & Treasurer Subject: Presenting the Treasurer’s Report for Fiscal Year 2025-2026 Q2 (Informational Item)
Date: February 18, 2026
Summary/Recommendation
Receive report on Ava’s cash position.

Background and Discussion
For the quarter ending December 31, 2025, Ava has maintained a positive cash balance on all Ava bank accounts. Below is a summary of account balances, cash received, and outstanding loan balances.

Account Balances as of 12/31/2025
Accounts Held Amount
River City Accounts
Operating
Internal Operating $ 2,982,223
Operating Fund $ 182,173,156
Lockbox $ 26,181,777
Interest Bearing
Money Market $ 60,773,088
Insured Cash Sweep Checking $ 90,302,021
Insured Cash Sweep Saving $ 5,163,661
CDARS $ 50,000,000
US Bank
Checking $ 58,237,429
Reserve Balance $ 330,873,400
Invested Capital $ 20,922,324
Wells Fargo
Security Margin $ 263,907
Total $ 827,872,986
Consent Item 6
Cash Received by Month into Lockbox Account
Outstanding Loan Balances:

PNC Credit Facility: $0.00
Customer Delinquency:

As of December 31, 2025
31 – 60 Days: $ 6,599,534
61 – 90 Days: $ 6,890,213
91 - 120 Days: $ 5,007,818
120+ Days: $ 55,477,688
Oct 2025 $ 125,126,144
Nov 2025 $ 86,248,379
Dec 2025 $ 87,475,914
Total $ 298,850,438
Consent Item 7
Consent Item 7
To: Ava Community Energy Authority
From: Matthew Chiodo – Senior Analyst, Power Resources
Subject: Approving Two Resolutions (1) Authorizing the Chief Executive Officer to Negotiate and Execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY9, LLC and (2) Authorizing the Chief Executive Officer to Negotiate and Execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY16, LLC
Date: February 18, 2026
__________________________________________________________________ Recommendation
(1) Adopt a Resolution authorizing the Chief Executive Officer to negotiate and execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY9, LLC. (2) Adopt a Resolution authorizing the Chief Executive Officer to negotiate and execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY16, LLC.

Background
In February 2024, the Ava Community Energy (“Ava”) Board of Directors (“Board”) approved the execution of five Power Purchase Agreements (“Agreements”) for solar projects located in Ava’s service territory with Prologis, Inc. (the “DAC Portfolio”). These solar projects are part of two California Public Utilities Commission (“CPUC”) funded programs: the Disadvantaged Communities Green Tariff (“DAC-GT”) and Community Solar Green Tariff (“CSGT”). These programs intend to promote development of renewable generation within Disadvantaged Communities (“DAC”) and offer primarily California Alternate Rates for Energy (“CARE”) and Family Electric Rate Assistance (“FERA”) program customers living in a DAC access to 100 percent renewable energy supply and a 20 percent discount on electricity bills.

Consent Item 7
The DAC Portfolio totals 7.28 MW of renewable energy from commercial rooftop solar projects in Ava’s service territory, including rooftop solar projects in Oakland, San Leandro, Hayward, and two projects in Tracy. This portfolio of long-term contracts is intended to serve approximately 3,700 customers in Ava’s service territory, consistent with the CPUC’s rules for the DAC-GT program1. Of the five projects, construction has begun on the Oakland, San Leandro, and Hayward projects, with the Oakland project expected to be online this spring.

Within the DAC Portfolio, the two Tracy projects, ESCA-PLD-TRACY16, LLC (“Tracy 2”) and ESCA-PLD-TRACY9, LLC (“Tracy 12”), have experienced interconnection delays during their development, making the development milestones currently unreachable, even with the allowed delay days in the Agreements. These challenges in meeting the online dates for Tracy 2 and Tracy 12 are due to delays in the interconnection process managed by Pacific Gas and Electric (“PG&E”). PG&E is the party responsible for managing the interconnection study process and syncing projects to the distribution system. Since the projects were unable to meet the interconnection deadlines outlined in the Agreements, the Milestone Dates, including the Expected Construction Start, Financial Close Date, Initial Synchronization Date, and Commercial Operation Dates, are no longer achievable. Therefore, contract amendments are needed to ensure that the projects remain viable and can come online later than originally planned.

When these two Tracy projects come online, they will provide valuable benefits to Ava customers, including bill savings for customers eligible for the Disadvantaged Community Green Tariff program. Both projects are expected to begin construction in late spring of 2026. Once completed, the projects will become the fourth and fifth DAC projects to be built and operational within Ava’s service territory. On November 18, 2024, Ava’s advice letter seeking additional eligible DAC-GT program volumes was approved by the CPUC, allowing Ava to increase the number of customers enrolled in the program. Currently, 4,196 Ava customers are enrolled in the DAC program. Staff will release a second request for offers (RFO) to solicit and contract additional DAC-GT program eligible long-ter contracts this spring. In the meantime, DAC-GT clean energy is served by short-term contracts. Below is a table with enrolled customers by city within Ava’s service territory:

City
# DAC-GT
Customers
Berkeley 33
Emeryville 81
Hayward 147
Lathrop 71
Newark 59
Oakland 1,925
1 CPUC D.18-06-027
Consent Item 7
San Leandro   323
San Lorenzo 1
Stockton 1,250
Tracy 217
Union City 89
Total 4,196
Attachments
A. Resolution of the Board of Directors of Ava Community Energy Authority Authorizing the Chief Executive Officer to Negotiate and Execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY9, LLC
B. Resolution of the Board of Directors of Ava Community Energy Authority Authorizing the Chief Executive Officer to Negotiate and Execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY16, LLC
Attachment Consent Item 7A
RESOLUTION NO. R-2026-XX

A RESOLUTION OF THE BOARD OF DIRECTORS

OF AVA COMMUNITY ENERGY AUTHORITY AUTHORIZING THE CHIEF

EXECUTIVE OFFICER TO NEGOTIATE AND EXECUTE A FIRST
AMENDMENT TO THE POWER PURCHASE AGREEMENT WITH ESCA-PLD-TRACY9, LLC WHEREAS Ava Community Energy Authority (“Ava”) was formed as a community choice aggregation agency (“CCA”) on December 1, 2016, under the Joint Exercise of Powers Act, California Government Code sections 6500 et seq., among the County of Alameda, and the Cities of Albany, Berkeley, Dublin, Emeryville, Fremont, Hayward, Livermore, Piedmont, Oakland, San Leandro, and Union City to study, promote, develop, conduct, operate, and manage energy-related climate change programs in all of the member jurisdictions. The Cities of Newark and Pleasanton, located in Alameda County, along with the City of Tracy, located in San Joaquin County, were added as members of Ava and parties to the Joint Powers Agreement (“JPA”) in March of 2020. The City of Stockton, located in San Joaquin County, was added as a member of Ava in September of 2022. The City of Lathrop, located in San Joaquin County, was added as a member to Ava in October of 2023. San Joaquin County was added as a member to Ava in July 2024. On October 24, 2023, Ava legally adopted the name Ava Community Energy Authority, where it had previously used the name East Bay Community Energy Authority since its inception; and
WHEREAS Ava issued the 2022 Disadvantage Communities Green Tariff and Community Solar Request for Offer in December 2022; and WHEREAS Ava executed a Power Purchase Agreement (“PPA”) with ESCA-PLD-TRACY9, LLC (“Seller”) on February 22, 2024; and WHEREAS The Power Purchase Agreement is for 3 mega-watts of Solar Energy with an Expected Construction Start Date of January 1, 2026, Financial Close Date of January 1, 2026, Initial Synchronization Date of March 1, 2026, and an Expected Commercial Operation Date of June 1, 2026; and
WHEREAS Ava staff is negotiating a first amendment to the PPA with the Seller to amend the Expected Construction Start Date, Financial Close Date, Initial Synchronization Date, and the Expected Commercial Operation Date.

NOW, THEREFORE, THE BOARD OF DIRECTORS OF AVA COMMUNITY ENERGY AUTHORITY DOES HEREBY RESOLVE AS FOLLOWS:

Section 1. The Chief Executive Officer, in consultation with staff and legal counsel, is hereby authorized to negotiate and execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY9, LLC to amend the Expected Construction Start Date, Financial Close Date, Initial Synchronization Date, and the Expected Commercial Operation Date.

Attachment Consent Item 7A
ADOPTED AND APPROVED this 18th day of February, 2026.

Betsy Andersen, Chair
ATTEST:

Adrian Bankhead, Clerk of the Board
Attachment Consent Item 7B
RESOLUTION NO. R-2026-XX

A RESOLUTION OF THE BOARD OF DIRECTORS

OF AVA COMMUNITY ENERGY AUTHORITY AUTHORIZING THE CHIEF

EXECUTIVE OFFICER TO NEGOTIATE AND EXECUTE A FIRST
AMENDMENT TO THE POWER PURCHASE AGREEMENT WITH ESCA-PLD-TRACY16, LLC WHEREAS Ava Community Energy Authority (“Ava”) was formed as a community choice aggregation agency (“CCA”) on December 1, 2016, under the Joint Exercise of Powers Act, California Government Code sections 6500 et seq., among the County of Alameda, and the Cities of Albany, Berkeley, Dublin, Emeryville, Fremont, Hayward, Livermore, Piedmont, Oakland, San Leandro, and Union City to study, promote, develop, conduct, operate, and manage energy-related climate change programs in all of the member jurisdictions. The Cities of Newark and Pleasanton, located in Alameda County, along with the City of Tracy, located in San Joaquin County, were added as members of Ava and parties to the Joint Powers Agreement (“JPA”) in March of 2020. The City of Stockton, located in San Joaquin County, was added as a member of Ava in September of 2022. The City of Lathrop, located in San Joaquin County, was added as a member to Ava in October of 2023. San Joaquin County was added as a member to Ava in July 2024. On October 24, 2023, Ava legally adopted the name Ava Community Energy Authority, where it had previously used the name East Bay Community Energy Authority since its inception; and
WHEREAS Ava issued the 2022 Disadvantage Communities Green Tariff and Community Solar Request For Offer in December 2022; and WHEREAS Ava executed a Power Purchase Agreement (“PPA”) with ESCA-PLD-TRACY16, LLC (“Seller”) on February 22, 2024; and WHEREAS the PPA is for 2 mega-watts of Solar Energy with an Expected Construction Start Date of January 1, 2026, Financial Close Date of January 1, 2026, Initial Synchronization Date of March 1, 2026, and an Expected Commercial Operation Date of June 1, 2026; and
WHEREAS Ava staff is negotiating a first amendment to the PPA with the Seller to amend the Expected Construction Start Date, Financial Close Date, Initial Synchronization Date, and the Expected Commercial Operation Date.

NOW, THEREFORE, THE BOARD OF DIRECTORS OF AVA COMMUNITY ENERGY AUTHORITY DOES HEREBY RESOLVE AS FOLLOWS:

Section 1. The Chief Executive Officer, in consultation with staff and legal counsel, is hereby authorized to negotiate and execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY16, LLC to amend the Expected Construction Start Date, Financial Close Date, Initial Synchronization Date, and the Expected Commercial Operation Date.

Attachment Consent Item 7B
ADOPTED AND APPROVED this 18th day of February, 2026.

Betsy Andersen, Chair
ATTEST:

Adrian Bankhead, Clerk of the Board

Consent Item 8
Consent Item 8
To: Ava Community Energy Authority
From: Matthew Chiodo – Senior Analyst, Power Resources
Subject: Approving Two Resolutions (1) Authorizing the Chief Executive Officer to Negotiate and Execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY9, LLC and (2) Authorizing the Chief Executive Officer to Negotiate and Execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY16, LLC (intentionally identical with Item 7)
Date: February 18, 2026
__________________________________________________________________ Recommendation
(1) Adopt a Resolution authorizing the Chief Executive Officer to negotiate and execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY9, LLC. (2) Adopt a Resolution authorizing the Chief Executive Officer to negotiate and execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY16, LLC.

Background
In February 2024, the Ava Community Energy (“Ava”) Board of Directors (“Board”) approved the execution of five Power Purchase Agreements (“Agreements”) for solar projects located in Ava’s service territory with Prologis, Inc. (the “DAC Portfolio”). These solar projects are part of two California Public Utilities Commission (“CPUC”) funded programs: the Disadvantaged Communities Green Tariff (“DAC-GT”) and Community Solar Green Tariff (“CSGT”). These programs intend to promote development of renewable generation within Disadvantaged Communities (“DAC”) and offer primarily California Alternate Rates for Energy (“CARE”) and Consent Item 8
Family Electric Rate Assistance (“FERA”) program customers living in a DAC access to 100 percent renewable energy supply and a 20 percent discount on electricity bills.

The DAC Portfolio totals 7.28 MW of renewable energy from commercial rooftop solar projects in Ava’s service territory, including rooftop solar projects in Oakland, San Leandro, Hayward, and two projects in Tracy. This portfolio of long-term contracts is intended to serve approximately 3,700 customers in Ava’s service territory, consistent with the CPUC’s rules for the DAC-GT program1
. Of the five projects, construction has begun on the Oakland, San Leandro, and Hayward projects, with the Oakland project expected to be online this spring.

City
# DAC-GT
Customers
Berkeley 33
Emeryville 81
Hayward 147
Lathrop 71
1 CPUC D.18-06-027
Consent Item 8
Newark 59
Oakland 1,925
San Leandro   323
San Lorenzo 1
Stockton 1,250
Tracy 217
Union City 89
Total 4,196
Attachments
A. Resolution of the Board of Directors of Ava Community Energy Authority Authorizing the Chief Executive Officer to Negotiate and Execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY9, LLC
B. Resolution of the Board of Directors of Ava Community Energy Authority Authorizing the Chief Executive Officer to Negotiate and Execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY16, LLC
Attachment Consent Item 8A
RESOLUTION NO. R-2026-XX

A RESOLUTION OF THE BOARD OF DIRECTORS

OF AVA COMMUNITY ENERGY AUTHORITY AUTHORIZING THE CHIEF

NOW, THEREFORE, THE BOARD OF DIRECTORS OF AVA COMMUNITY ENERGY AUTHORITY DOES HEREBY RESOLVE AS FOLLOWS:

Attachment Consent Item 8A
ADOPTED AND APPROVED this 18th day of February, 2026.

Betsy Andersen, Chair
ATTEST:

Adrian Bankhead, Clerk of the Board
Attachment Consent Item 8B
RESOLUTION NO. R-2026-XX

A RESOLUTION OF THE BOARD OF DIRECTORS

OF AVA COMMUNITY ENERGY AUTHORITY AUTHORIZING THE CHIEF

NOW, THEREFORE, THE BOARD OF DIRECTORS OF AVA COMMUNITY ENERGY AUTHORITY DOES HEREBY RESOLVE AS FOLLOWS:

Section 1. The Chief Executive Officer, in consultation with staff and legal counsel, is hereby authorized to negotiate and execute a First Amendment to the Power Purchase Agreement with ESCA-PLD-TRACY16, LLC to amend the Expected Construction Start Date, Financial Close Date, Initial Synchronization Date, and the Expected Commercial Operation Date.

Attachment Consent Item 8B
ADOPTED AND APPROVED this 18th day of February, 2026.

Betsy Andersen, Chair
ATTEST:

Adrian Bankhead, Clerk of the Board

CEO Report Item 9
CEO Report Item 9
TO: Ava Community Energy Board of Directors
FROM: Howard Chang, Chief Executive Officer
SUBJECT: CEO Report (Informational Item)
DATE: February 18, 2026
________________________________________________________________________ Recommendation
Accept Chief Executive Officer (CEO) report on update items below.

Executive Committee Meeting
The February 4, 2026 Executive Committee Meeting was canceled. The next Executive Committee meeting will be held on Wednesday, March 4, 2026 at 3pm.

Financial, Administrative and Procurement Subcommittee Meeting A Financial, Administrative and Procurement Subcommittee Meeting was held on Wednesday, January 28, 2026. Members received a preview of the FY 2025-26 Mid-Year Budget Review and an overview of Contracts Entered Into. The next Financial, Administrative and Procurement Subcommittee meeting will be held on Wednesday, March 25, 2026 at 3:30pm.

Unincorporated San Joaquin County Enrollment
Ava will begin providing electricity generation service to unincorporated San Joaquin County’s 60,000 customers starting in May 2026. Most customers will start service in May and will see Ava on their PG&E bill starting in June 2026. Customers with solar will be enrolled closer to their true-up date, with enrollment occurring throughout the year.

Ava is required by the California Public Utilities Commission (CPUC) to send a total of four customer notifications, two notifications prior to enrollment and two notifications following enrollment. Mailers will be sent via mail, and email, when available.

CEO Report Item 9

Images of notices that will be sent to new customers in unincorporated San Joaquin County, starting in March 2026
Marketing & Outreach
To further ensure that these new customers are aware that their electricity generation provider is changing, Ava’s Marketing team will begin outreach at the end of February, starting with digital ads to raise awareness before notices go out. Additional marketing efforts during the enrollment period will include:
• Billboards, streaming TV, and print ads
• Targeted micro-campaigns to reach specific neighborhoods • Community partnerships with municipal partners, agricultural and business organizations, and other trusted community leaders
• Social media, newsletters, and local press
• Webinars and in-person presentations
Ava in the Community - February
To raise awareness of our service, support our community, and stay close to our customers, Ava sponsors and staffs events throughout our service territory all year long. Stop by to say hello at any of these upcoming events between 02/18 and 03/18. Learn more at https://avaenergy.org/about-ava/meetings-events/events/ Date Event Location
3/14/26 - 3/15/26 Dublin St. Patrick’s Day Dublin
3/14/26 Oakland Roots vs New Mexico United Oakland Coliseum Evaluating the Drivers of PG&E’s Electricity Rate Increases California consistently ranks among the top three states with the highest residential electricity rates in the nation. Adjusted for inflation, PG&E’s average electricity rate increased by approximately 49% from 2019 to 2024, followed by a decrease of 16% from 2024 to 2026. Rate increases drive household electricity bills higher, worsening affordability challenges for families already burdened by high housing, healthcare, and food costs. High rates also threaten the CEO Report Item 9
state’s clean energy transition by eroding the financial advantages customers receive from electrification.

Although rising rates are largely driven by the costs of adapting to climate impacts like extreme wildfires, rising electricity bills may make the public more receptive to claims blaming California’s climate policies for high energy costs. This study makes clear that the largest driver of recent electricity cost, rate, and bill increases in California is responding to climate change (e.g., wildfire mitigation) rather than preempting it (e.g., clean energy procurement).

To investigate the drivers of rising electricity rates, Ava Community Energy analyzed PG&E costs and rates as published in state and federal regulatory filings. Read the full report for Ava’s key findings and recommendations for helping to address the identified challenges. https://avaenergy.org/insight/report-evaluating-the-drivers-of-pge-electricity-rate-growth/ Local Development Update
Given the high number of local development activities and varied status of design, implementation, and operations, we will aim to provide a summary update of all programs and services on a quarterly basis via the CEO report. The purpose is to keep the Board appraised of Local Development Programs and activities at a summary level. Action items and deep dives will continue to be brought forth on an as needed basis.

CEO Report Item 9
Program Name Status Description Key Metrics Key Upcoming Dates Planning
Customer
Programs
Roadmap On Hold
Define Ava's customer
programs roadmap for 2027-
2032 N/A
Ava evaluating options to
progress the Roadmap
Transportation
Electrification
Ava Charge Operational
Developing EV fast charging
sites on City lots and garages
across Ava's service territory
with a focus on locating near
multi-family dwellings
Oakland City Center West
garage with 31 chargers is now
in operation
Ava negotiating with developers
currently under contract to
deliver the next four Ava Charge
projects
Ava SmartHome
Charging Operational
Program to shift home EV
charging away from peak
hours while helping customers
lower their cost of home EV
charging.
>2,700 EVs enrolled, Virtual
Power Plant performance
metrics are being met
Refining customer engagement
and dispatch strategy to
increase participation and
optimize performance
Ava Bike Electric Operational
Program to provide incentives
for customers to purchase
electric bikes, CARE
customers receive higher
incentives and 40% of
allocated funds
over 19,000 applications
received, 2,746 e-bikes
purchased from 29 participating
bike shops. CARE customers
have redeemed 51% of rebates
Drawing on 4th Tuesday of each
month.

Building
Electrification
Health-e
Communities Pilot Operational
Direct installation program to
replace gas stoves with
induction for CARE customers
and complete indoor air quality
monitoring to evaluate air
quality impacts. Installed 162 induction cooktops
Ava is reviewing the results of
the pilot and will come back to
the Board with a
recommendation. Applications
closed in September
CEO Report Item 9
Building Efficiency
Accelerator Operational
CPUC funded program to
complete energy efficiency
and building electrification
projects on commercial and
industrial customers, like LED
lighting and heat pump water
heaters
On target to reduce
consumption by 12 GWh, 14
customers engaged in 100
unique projects
Program enrollment ends April
2026
SolarApp Heat
Pump Water
Heater (HPWH)
permitting Canceled
Contract with SolarApp to add
heat pump water heaters to
their online permitting
application
SolarApp will deliver lessons
learned from permitting
departments, HPWH installers,
and manufacturers on the value
of expedited permitting
Contract cancelled do to
Federal withdrawal of IRA
funding and SolarApp needing
to focus on core solar permitting
business
Resilience
Critical Municipal
Facilities
In
Development
Program to provide resilience
on critical municipal sites.

Eight projects across four
cities will provide bill savings
Installing 258kW of PV and

1.34MWh of batteries
generating 778,000 kWh of
clean energy in year 1.

Permits received on target and
construction started on first
Livermore project. Ava is on
target to achieve mechanical
completion on NEM 2.0 projects
by the April 15, 2026 deadline.

SmartHome
Battery +
Resilience Hubs
In
Development
Program to provide $15M in
upfront and ongoing incentives
for residential solar and
storage projects and
community resilience hubs
Manage 21MW of solar paired
batteries for residential and low
income customers and
community resilience hubs.

SmartHome Battery Program
launch planned in March.

Resilience Hub technical
assistance launching in Q2 '26
Community
Grants
Community Grant
Program Operational
$7.05M budgeted for
Community Grants to work
with Community Based
Organizations across four
strategic focus areas Award ~24 grants through 2033
Staff hosted Community Grants
public workshop to solicit ideas
for next round of community
grants. Staff compiling results
and will develop next round of
grant RFPs for issuance in 2026
with community input
CEO Report Item 9
EV Charging Operational
Both parties have identified
multiple sites and are
evaluating feasibility studies.

Two grants of $300k over 3
years
Green the Church has installed
their first project at Glad Tidings
Church in Hayward
Interfaith Power and Light and
Green the Church continue to
work on identifying sites to build
community-owned level 2 EV
charging stalls.

Youth Training Operational
Empowering young individuals
through education and
exposure to employment
opportunities and hands-on
training in clean energy
technologies through
partnerships with Rising Sun
Center for Opportunity and
AGAPE (a partnership
between Cypress Mandela
Training Center and
Revalue.io)
Two grants of $300k over 3
years
trained over 150 youth in
electrification trades
Students are beginning to
matriculate into the job market.

Rising Sun is developing an
online jobs platform to help
graduates find work in the
trades.

Community
Resilience Hubs Operational Emerald Cities Collaborative Grant of $300k over 1.5 years,
completed three community
workshops with over 170
participants and 26 unique
community based organizations
ECC delivered first of three
Community Workshops to
engage potential Resilience Hub
sites. Will continue outreach per
the grant agreement
Staff Report Item 11
Staff Report Item 11
To: Ava Community Energy Authority Board of Directors
From: Howard Chang, CEO
Subject: Mid-Year Budget Review
Date: February 18, 2026
Recommendation
Receive the report informing Ava’s Board of Directors (the “Board”) of the mid-year budget performance for the 2025-26 fiscal year.

Analysis and Context
Every June, the Board approves the budget for the following fiscal year spanning from July 1st of that calendar year through June 30th of the following calendar year. Staff internally reviews budget performance regularly during the fiscal year and presents an update of these results to the Board at the mid-year mark.

Ava’s accounting generally requires about 45 days to close the preceding period to ensure all revenues and expenses are captured accurately. As such, the close of December 31st is typically mid-February. Presenting the mid-year analysis in February allows a more accurate presentation on the half-year performance.

The mid-year report presented covers the closed months of July through December 2025 actuals, and updated projections for the second half of the fiscal year: January through June 2026. Below is a concise high-level summary of our year-to-date financial performance.

Revenues
Revenues from sales are expected to come in approximately $42MM above budget for the fiscal year due to moving forward the timeline to incorporate San Juaquin County from November 2026 to May and Board approved changes to the value proposition in calendar year 2026.

These changes improved revenues by $7.8MM and eliminated the need to use the budgeted $35 million from the Rate Stabilization Fund (RSF).

Staff Report Item 11
Cost of Energy
Total cost of energy is down by about $19.3MM driven predominately by favorable energy market conditions and reduction in renewable targets for 2026 Overhead
Total overhead expenses are expected to be down by about $3.5MM due primarily to a more measured approach in filling vacant positions throughout the year.

Non-Operating Activity
Total annual net non-operating activity is expected to be about $13.9MM above budget due to “higher for longer” interest rates on invested capital and reductions to Local Development Net Revenues
Total annual net revenue is expected to be $44.5MM above budget without the utilization of RSF.

Staff Report Item 11
Attachment
A. Power Point Presentation to Board
Fiscal Year 2025-26 Mid-Year Budget to Actuals
Review
Board of Directors
Howard Chang | February 18, 2026
Attachment Staff Report Item 11A
Introduction
2 • Every June, Ava’s Board of Directors (the “Board”) approves spending for the following fiscal year spanning from July 1st through June 30th of the following year. • Staff reviews budget performance internally regularly and presents these results to the Board at, or about, the mid-year mark.
• Ava’s accounting cycle generally requires about 45 days to close a period to ensure all revenues and expenses are captured accurately. Thus, the close of December 31st is typically mid-February. • Presenting the mid-year analysis in February allows a more accurate presentation on the true half year performance.
• The mid-year report presented today covers the closed months of July through December, and updated projections for the second half of the fiscal year—January through June. These updated values will collectively be referred to as “updates”
• This mid-year budget update incorporates more substantive updates from a typical year due to significant rate changes and following the board actions taken in December 2025 Attachment Staff Report Item 11A
Ava Historical and 2026 Residential Generation Rates
3 • PG&E's final Annual Energy True-up (AET) was released on December 31st, which finalized 2026 rates
▪ Avg 2026 rates in the AET were higher than the ERRA by ~12% ▪ Relative to the Oct ERRA, this results in a 1.3 cent increase in Bright Choice customer rates with 0.7 cents from rates and

0.6 cents in PCIA
• Ava reduced 2026 Bright Choice generation rates by 42% (exclusive of PCIA) and 20% (inclusive of PCIA)
▪ The Dec update reflected a 29% reduction (inclusive of PCIA) • For Renewable 100 , 2026 generation rates are reduced by 34% (exclusive of PCIA) and 14% (inclusive of PCIA)
▪ The Dec update reflected a 21% reduction (inclusive of PCIA) • Average TOU-C residential Bright Choice customer bill is ~ $10 less per month compared to 2025
• Average TOU-C residential Renewable 100 customer bill is ~ $8 less per month compared to 2025
Note: All data on this slide reflects 2018 PCIA vintage TOU-C (Ava’s most common residential rate)
Attachment Staff Report Item 11A
2026 Forward Energy Market Volatility
4 • The graph shows the forward price for full calendar year 2026 at the NP15 node over the past 5 years. NP15 is a close proxy to PG&E DLAP. Prices have been highly volatile for the past 5 years, reaching a peak in 2023 and have since steadily softened.

Attachment Staff Report Item 11A
Historical Energy Market Volatility
5 • The graph to the right shows historical daily NP15 energy and PG&E Citygate natural gas hub prices.

Attachment Staff Report Item 11A
Price Volatility—Impact on Market Prices to Fiscal Year Outcomes 6 • Historically, March through May are
low price volatility months
• Stressed price scenarios are for
February through June
• +/-1 SD move across the 6-month
period results in a $12.7MM
increase/decrease in energy costs
• +/-2 SD move across the 6-month
period results in a $25.1MM
increase/decrease to energy costs
Chart shows expected market prices (mean) and +/-1 and +/-2 standard deviation moves by calendar month
Attachment Staff Report Item 11A
7 Ava Community Energy
Fiscal Year 2025-2026 Mid-Year Budget Update
• Expecting increase in operating revenue by $7.8MM
▪ Operating Revenue increases by $42MM without use of
Rate Stabilization Fund (GASB 62)
• Cost of Energy expected to improve by
$19.2MM based on favorable energy market
conditions
• Overhead expenses forecasted to be reduced by
$3.5MM
• Non-Operating activity is expected to add an
addition $13.9MM to net revenues
• Net Revenues are projected to be positive by about
$44.5MM
▪ Net Revenue increases by $79.8MM without use of
Rate Stabilization Fund (GASB 62)
The following slides provide more detail on these
changes
Attachment Staff Report Item 11A
8 First Half FY2026 Budget to Actuals
• Revenues from sales were $37.7MM under
budget driven by lower load resulting from milder
than expected weather
• Cost of Energy was $23.4MM under budget driven
by lower load/lower energy prices, higher than
forecasted RA sales, and higher than forecasted
large hydro production. This was partially offset
by higher REC costs from higher than expected
market price benchmarks.
• Overhead expenses are $6.3MM below budget
due to cost reductions, deferred hiring, and
deferred functional area costs
• Non-Operating activity increased net revenues by
$11.2MM from higher interest rates on
investments and reduced Local Development
funding from the December Board decision
• Net revenues were $5MM above expected budget
Attachment Staff Report Item 11A
9 Second Half FY2026 Original Budget to Updated
Budget
• Revenues from electricity sales is forecasted to
increase by $43.6MM driven by changes in value
proposition, higher than forecasted 2026 rates, and SJ
County enrollment in May
o Operating Revenue increases by $79.7MM when excluding RSF (GASB62)
• Cost of Energy is forecasted to increase by $4MM
relative to budget driven by higher REC costs due to
increased market price benchmarks and partially
offset by lower energy costs
• In Overhead, continued savings from deferred hiring
for open positions while deferred costs from the first
half in other functional areas are back half weighted
• Non-Operating activity is expected to add about
$2.7MM due to reduced Local Development Funding
offset by reduced expected interest income
• Net Revenues are expected to increase by $39.5MM
from the original budget
o Net Revenues increase by $74.9MM without use of RSF
GASB62
Attachment Staff Report Item 11A
10 Second Half Calendar 2026 Dec BOD to Updated
• Revenues from electricity sales is forecasted to
increase by $73.9MM driven by changes in value
proposition and higher than December estimated
2026 rates
• Cost of Energy is forecasted to decrease by
$42.4MM driven by $14.3MM in lower energy
prices, $11MM additional sales of RA length, and
$17MM reduced REC costs due to Board action
and lower market prices
• No expected changes in Overhead
• Non-Operating Income reduced by expected
changes in interest rates
• Non-Operating Expenses includes reduced Local
Development Funding
• Net Revenues are expected to increase by $114MM
from December estimates, but still retains a
structural deficit of $87.5MM
Attachment Staff Report Item 11A
11 Calendar 2027 Preliminary
Forecast
• This forecast is indicative and
not board approved
• Maintains the adjusted value
proposition of 0.5% discount on
BC and 1¾ ¢ on R100 and has $0
of incremental local
development funding
• Relative to the December Board
update, the Net Revenues are
reduced by ~$100MM when
accounting for the value
proposition adjustments
• This forecast does not include
any changes that may occur
resulting from PCIA Track 2 or 3
o The CPUC initiated Track 2 in early
2026
• The RSF balance is forecasted
to be higher relative to the
December forecast driven by
board actions, higher rates,
and lower forecasted energy
costs
o This is partially offset by lower
net revenues in 2027
• Staff will continue to update
projections going into the
May/June budget setting
process for FY 2026-27
• FY 2025-26 results will be
discussed with the financial
audit in fall 2026
Attachment Staff Report Item 11A
Staff Report Item 12
Staff Report Item 12
To: Ava Community Energy Authority
From: Arielle Romero Cox, Director of Strategic Load
Subject Overview of large electric load growth trends and implications for Ava Community Energy
Date: February 18, 2026
__________________________________________________________________ Summary/Recommendation
This item is informational only and no action is required.

This Staff Report provides an overview of forecasted large electric load growth driven particularly from electric vehicles and data centers. These trends may have long-term implications for Ava’s procurement, rates, and reliability planning.

Analysis and Context
Statewide Load Growth: A Shift from Historic Trends
After nearly two decades of relatively flat electricity demand, California is entering a period of sustained load growth. Megawatts (MW) are a standard unit for measuring electricity power. On January 21, 2026, the California Energy Commission (CEC) adopted the California Energy Demand Forecast, 2025-2045. The CEC forecast predicts a significant rise in the State’s peak electricity demand by 2045, potentially increasing by 42% (to 66,000 MW low estimate), 53% (69,000 MW mid estimate), up to 61% (74,900 MW high estimate), driven by increases in electric vehicles, data centers, and building electrification (fuel shifting). This represents a structural shift in California’s electricity system and planning assumptions, with implications for generation procurement, grid infrastructure, and cost allocation.

Staff Report Item 12
Primary Drivers of Demand Growth in California
The Integrated Energy Policy Report (IEPR), by the CEC, and related analyses indicate that load growth is not evenly distributed across end uses. The CEC’s 2025 IEPR identifies three primary drivers of future peak demand growth within the CAISO system: electric vehicles adding approximately 8,234 MW, baseline consumption growth driven by economic and demographic trends adding 6,011 MW, and data centers adding 4,721 MW between 2025 and 2045. While this growth is distributed statewide and will not be evenly realized across service territories, the combined scale of EV and data center–driven peak demand alone (nearly 13,000 MW) is more than seven times Ava’s 2025 system peak of 1,657 MW. Other Loads: Industrial electrification and hydrogen production contribute comparatively less to total demand growth in the IEPR forecast, particularly within the next 10–15 years.

It is important to distinguish between peak load and total annual energy demand, as each affects planning and cost outcomes in different ways. Peak load reflects the system’s maximum instantaneous demand and drives infrastructure sizing, resource adequacy, and capacity procurement, while total annual demand reflects cumulative energy consumption over time and drives overall power procurement costs. A useful analogy is a highway system. Think of peak load as the number of freeway lanes needed during rush hour, whereas total annual demand reflects the total volume of traffic using the freeway throughout the year. For example, electric load from an electric vehicle charging hub would increase “rush-hour” demand due to concentrated charging behavior, while data centers operate more like freight traffic running continuously, contributing more to total energy consumption. As a result, data centers can have Staff Report Item 12
an outsized impact on Ava’s long-term procurement requirements and rate design, even when their contribution to peak demand is smaller than that of EVs.

This concentration of growth among a small number of load types heightens planning risk if forecasts do not align with actual project development timelines.

Why is Data Center Growth Accelerating in California?
Data center growth in California, particularly in urban regions like Ava’s service territory, is being driven by several factors:

Proximity to Major Population and Technology Hubs: Ava’s service area sits within one of the largest population and economic centers in the state, where demand for latency-sensitive applications such as AI, cloud computing, streaming, healthcare, and public-sector services is concentrated. Proximity to end users and technology firms improves performance and supports advanced digital workloads.

Strong Grid Connectivity and Clean Energy Access: The region benefits from established transmission infrastructure and access to clean energy resources, which are increasingly critical site-selection requirements for data center developers and tenants with reliability and decarbonization commitments.

Capital and Market Dynamics: Continued growth in AI and cloud services, combined with strong investor interest, is driving development in locations that offer long-term market stability and high-value demand, even where power and land costs are higher than in other states.

Staff Report Item 12
California Relative to Other U.S. States
California is a high-cost data center market in the U.S. It does not compete directly with states such as Virginia, Texas, or Arizona on power prices or land costs. New data centers outside California are getting significantly larger, with average sizes shifting from historical 30-50 MW to current major projects often hitting 100-200 MW+ per building, and many major new developments targeting 500 MW to over 1 GW (1000 MW) of total capacity, especially in markets like Northern Virginia, Dallas-Fort Worth, and Phoenix, driven by AI demand. • Northern Virginia dominates hyperscale data center development due to low-cost power, favorable tax treatment, and proximity to federal government workloads. • Texas and Arizona are experiencing rapid growth due to inexpensive land, fast permitting, and lower electricity costs.

Total Inventory vs. Under Construction by Primary Market, H1 2025 Potential Scale Within Ava’s Service Territory
In California, data center development has been concentrated in Santa Clara and Los Angeles, due to proximity to fast fiber-optic connections and lower electricity rates served by those cities’ municipal utilities. More recently, Pacific Gas & Electric has received a surge of requests for service to new data centers. Based on independent market analysis, Ava’s total addressable Staff Report Item 12
data center market is expected to reach 530 MW–770 MW peak demand by 2030, which represents roughly 11–16% of the CEC’s 2025 IEPR–forecasted statewide data center peak growth of approximately 4,721 MW between 2025 and 2045. While not all forecasted statewide growth will materialize in specific service territories, this potential share reflects strong demand conditions in and around the San Francisco Bay Area and Silicon Valley, where developers continue to pursue new capacity despite grid and land constraints. Silicon Valley remains one of the nation’s most active markets, with notable strategic developments underway such as the City of San José–PG&E Transmission Agreement to deliver up to 2,000 MW of new transmission capacity and Silicon Valley Power’s (City of Santa Clara) $459 Million+ System Expansion Program to double electric load capacity to 1,300 MW in 5-10 years. These local trends underscore why even a modest share of statewide data center growth in Ava’s territory could be meaningful for planning and procurement.

Most new data centers seeking to build in California are sized between 50–99 MW.

While not all announced or proposed projects materialize, the existing proposed data center project pipeline in Ava’s service territory of 198+ MW indicates that Ava’s service area is an attractive region for future large-load development. One megawatt of data center capacity running at an 80% load consumes enough electricity to power approximately 1,300 average homes in Oakland for a full year.

Types of Data Centers
There are four key data center types that differ by their scale and ownership models.

Edge data centers provide real-time data processing - in applications such as autonomous vehicles, streaming services, and gaming - and are located close to end-users to minimize latency (a measure of how quickly data can be transmitted and processed). Typical size <1 MW.

Enterprise data centers are owned, operated, and maintained by a single company, and are designed specifically to support the specific IT needs of the owner's organization. They are typically located at an existing company site and perform tasks such as storage of sensitive data and hosting internal applications. Typical size 1-5 MW.

Colocation data centers are facilities owned and operated by a single company, and lease physical space to multiple tenant companies for housing their own IT infrastructure. These facilities are typically located in areas with access to multiple communications providers and robust connectivity options, offering redundancy and flexibility for customers. Typical size 5-20 MW.

Hyperscale cloud data centers are large-scale facilities designed, built, and operated by a single organization to support uses such as AI model training and running of cloud computing platforms (e.g. Google Cloud). Sites are typically in locations with access to low-cost power, affordable land, and strong network connectivity. Typical size 20-100+ MW.

Staff Report Item 12
Managing the Large-Load Development Pipeline
Ava does not control land use approvals or electric interconnections; those functions are led by local jurisdictions and PG&E, respectively. However, the timing, scale, and certainty of large-load development directly affect Ava’s generation procurement, resource adequacy, and rate-setting responsibilities. As a result, Ava has a strong interest in early visibility into potential large-load projects, even though it does not approve or construct them.

In November 21, 2024, PG&E filed an application at the CPUC to establish a streamlined approach for energizing new transmission-level retail customers called Electric Rule 30. While the application is broad in its definition of “large load customers” that would apply for service under this tariff, the California Public Advocates Office (PAO) notes 75% of PG&E customer requests to interconnect at the transmission level in the last two years are from data centers1 .

The Electric Rule 30 proposal is solely focused on the energization costs, not a unique large load rate. Under the proposed Rule 30, large load customers are required to pay upfront the cost of new interconnection facilities, or the customer can build the facilities themselves and transfer ownership to PG&E, if more cost-effective. Costs for upgrades on the existing transmission system would be paid by ratepayers. PG&E states this is due to these facility upgrades benefitting multiple customers as well as advancing California policy initiatives such as electrification. Rules 30 also includes information sharing provisions to facilitate community choice aggregators (CCAs) and PG&E having early notice of potential large load customers and being able to make cost effective generation resource decisions.

To better align planning assumptions with real-world outcomes, Ava will execute a multi-pronged large load customer engagement approach, including:
• Direct engagement with developers early in the project lifecycle to understand project maturity, timing, and likelihood of completion.
• Coordination with city and county planning and economic development staff, who have visibility into zoning, permitting, and development priorities. • Ongoing information sharing with PG&E regarding interconnection activity, infrastructure constraints, and expected energization timelines.

Through this approach, Ava seeks to ground its load forecasts and procurement planning in qualified, credible project information rather than preliminary or speculative requests, reducing the risk of misaligned energy procurement or rate design. Board awareness and engagement at the local level can help reinforce these partnerships and ensure Ava is positioned as a trusted technical resource for member agencies as they consider large-load development opportunities.

Ratepayer Protection and Clarifying Utility Roles
Community concerns related to data centers and other large loads often focus on potential 1 Pg. 2, “PAO Opening Brief,” A. 24-11-007, filed October 24, 2025.

Staff Report Item 12
impacts on electricity rates and environmental outcomes. These concerns arise in the context of rising statewide electric rates, which are driven largely by delivery-related costs specifically wildfire mitigation, infrastructure hardening, and legacy investments administered by PG&E as the electric distribution utility. Despite being one of the top data center markets in the country, California has not seen an increase in electricity rates due to data center growth, according to the California Public Utilities Commission (CPUC)2
.

PG&E is responsible for delivering electricity, including operating and maintaining poles, wires, substations, and interconnections, while Ava is responsible for procuring electricity and related commodities on behalf of customers. Ava does not control delivery rates or infrastructure investment decisions, but it does control how electricity is procured and ratemaking for generation.

As Ava evaluates future approaches to serving large-load customers, a core guiding principle will be ensuring that the costs and risks associated with substantial energy consumption, particularly generation and procurement-related costs are borne by the customers driving that load, rather than being shifted to residential and small business customers. This principle aligns with Ava’s dual objectives of maintaining rate stability while supporting responsible economic development within its service territory.

Benefits for Ava and Its Member Communities
When managed with appropriate safeguards, large electric loads may present strategic opportunities for Ava and its member communities. Incremental, creditworthy load growth can help spread fixed administrative and program costs over a larger sales base and, in some cases, support long-term clean energy procurement at greater scale provided that the costs and risks associated with that load are fully borne by the customers driving it. Large loads can also enable investments in new clean resources and grid infrastructure that support statewide decarbonization goals, and the digital services relied upon by residents, businesses, and public agencies. Ava’s interest at this stage is not to promote or discourage any specific development, but to ensure that future decisions protect existing customers while responsibly managing growth within its service territory. Ava’s staff want to work with our member cities/counties to understand how they view data centers and whether they want to attract such development.

In addition to grid impacts, large data center developments are often associated with meaningful economic activity at both the state and local levels. PG&E has cited estimates indicating that each 1 gigawatt of new data center load could support approximately 5,000 construction jobs, 500 permanent technology jobs, more than 11,000 construction-related support jobs, and nearly 3,000 permanent associated jobs, while also generating billions in annual statewide economic 2 Utilities and Energy: Joint Oversight Hearing on AI's Energy Impacts Background Report

1.28.2026
Staff Report Item 12
activity, including $125–$175 million in incremental property tax revenue and $250–$300 million in additional sales tax revenues.
3 At the local level, several Ava member jurisdictions also levy a Local Utility Users Tax (UUT) on electricity consumption ranging from 3.25-9.5%. Large continuous loads such as data centers can provide a direct and ongoing revenue stream to host cities to support essential services and community investments. While these figures represent statewide estimates and outcomes will vary by location, they highlight the potential economic benefits that can accompany thoughtfully planned large-load development.

Looking Ahead
As large-load development proposals become more concrete, staff may return to the Executive Committee and Board with future informational updates or action requests. These may include discussions of large-load tariffs or non-standard rate agreements, designed to improve load forecasting, address near-term reliability needs, and ensure cost causation principles are upheld.

Any future proposals would be brought forward separately for Board consideration, with opportunities for policy discussion and public input.

Attachments
A. PowerPoint Presentation
B. CEC adopted California Energy Demand Forecast, 2025-2045 C. Utilities and Energy: Joint Oversight Hearing on AI's Energy Impacts Background Report

1.28.2026
3 https://investor.pgecorp.com/news-events/press-releases/press-release-details/2025/PGE-Data-Center-Demand-Pipeline-Swells-to-10-Gigawatts-with-Potential-to-Unlock-Billions-in-Benefits-for-California/default.aspx Overview of large electric
load growth trends and
implications for
Ava Community Energy
Arielle Romero Cox, Director of Strategic Load
February 18, 2026
Attachment Staff Report Item 12A
Navigating California's Energy Future
This report provides an overview of the forecasted large electric load growth, primarily driven by electric vehicles (EVs) and data centers.

These trends have significant long-term implications for Ava Community Energy's procurement strategies, rate structures, and reliability planning. This information is for awareness only; no immediate action is required.

Attachment Staff Report Item 12A
California's Shifting Energy Landscape
After nearly two decades of stable electricity demand, California is now entering a period of sustained load growth. The California Energy Commission (CEC) projects a substantial increase in the state's peak electricity demand by 2045. 1 Projected Peak Demand Increase
The CEC's forecast predicts a 42% to 61%
rise in peak electricity demand by 2045,
reaching up to 74,900 MW.
2 Key Growth Drivers
This surge is primarily fueled by electric
vehicles, data centers, and building
electrification (fuel shifting).
3 Structural Shift
This represents a fundamental change in
California's electricity system, impacting
generation procurement, grid infrastructure,
and cost allocation.

Attachment Staff Report Item 12A
Attachment Staff Report Item 12A
Projected Peak Electricity Demand Growth by 2045
Attachment Staff Report Item 12A
Primary Drivers of Demand Growth
The Integrated Energy Policy Report (IEPR) highlights that load growth is not uniformly distributed across all end uses. Three main drivers are identified for future peak demand growth within the CAISO system between 2025 and 2045.

Electric Vehicles (EVs)
EVs are projected to add
approximately 8,234 MW to the
peak demand.

Baseline Consumption
Economic and demographic
trends contribute an additional
6,011 MW.

Data Centers
Data centers are expected to add
4,721 MW to the peak demand.

The combined peak demand from EVs and data centers alone (nearly 13,000 MW) is more than 7x Ava’s 2025 system peak of 1,657 MW Attachment Staff Report Item 12A
Peak Load vs. Total Annual Energy Demand
Peak Load
Peak load represents the system's maximum
instantaneous demand. It dictates
infrastructure sizing, resource adequacy,
and capacity procurement. Think of it as the
number of freeway lanes needed during rush
hour.

Total Annual Demand
Total annual demand reflects cumulative
energy consumption over time, driving
overall power procurement costs. This is
analogous to the total volume of traffic using
the freeway throughout the year.

For example, EV charging hubs increase "rush-hour" demand due to concentrated charging, while data centers operate continuously, contributing more to total energy consumption. Data centers can significantly impact Ava's long-term procurement and rate design.

Attachment Staff Report Item 12A
Why Data Center Growth is
Accelerating in California
Data center growth in California, particularly in urban areas like Ava’s service territory, is driven by a confluence of strategic advantages.

Proximity to Hubs
Close to major
population and
technology centers,
meeting demand for
latency-sensitive
applications like AI,
cloud computing, and
streaming.

Strong Grid & Clean
Energy
Established transmission
infrastructure and
access to clean energy
are critical for
developers with
reliability and
decarbonization
commitments.

Capital & Market
Dynamics
Growth in AI and cloud
services, coupled with
strong investor interest,
drives development in
stable markets with
high-value demand.

Source: DataCenters.com
Attachment Staff Report Item 12A
California's Data Center Market: A Unique Landscape
California stands out as a high-cost data center market compared to other U.S. states, not directly competing on power prices or land costs with regions like Virginia, Texas, or Arizona.

New data centers outside California are significantly larger, with average sizes shifting from 30-50 MW to major projects often exceeding 100-200 MW, and many targeting 500 MW to over 1 GW of total capacity, especially in markets like Northern Virginia, Dallas-Fort Worth, and Phoenix, driven by AI demand.

Northern Virginia: Dominates hyperscale development due to low-cost power, favorable tax treatment, and proximity to federal government workloads.

Texas and Arizona: Experiencing rapid growth thanks to inexpensive land, fast permitting, and lower electricity costs.

Total Inventory vs. Under Construction by Primary Market, H1 2025 Attachment Staff Report Item 12A
Potential Scale Within Ava’s Service Territory
While California's data center development has historically concentrated in Santa Clara and Los Angeles, Ava's service territory is poised for significant growth.
1 Market Potential
Ava’s total addressable data center market is projected to reach 530 MW–770 MW peak demand by 2030, representing 11–16% of the CEC’s statewide forecast.
2 Strong Regional Demand
This reflects robust demand in the San Francisco Bay Area and Silicon Valley, where new capacity is pursued despite grid and land constraints.
3 Local Initiatives
Strategic developments like the San José–PG&E Transmission Agreement (2,000 MW) and Silicon Valley Power's expansion (1,300 MW) underscore the region's growth.
4 Project Pipeline
Ava’s service area has an existing proposed data center project pipeline of 198+ MW, indicating its attractiveness for large-load development.

One megawatt of data center capacity running at 80% load can power approximately 1,300 average homes in Oakland for a full year.

Attachment Staff Report Item 12A
Understanding Data Center Types
Data centers vary in scale and ownership models, each serving distinct purposes and contributing differently to energy demand.

The Corner Store
Edge Data Centers
Located close to end-users for
real-time data processing (e.g.,
autonomous vehicles, streaming).

Typical size: <1 MW.

The Home Owner
Enterprise Data Centers
Owned and operated by a single
company for specific IT needs (e.g.,
sensitive data storage, internal
applications). Typical size: 1-5 MW.

The Landlord
Colocation Data Centers
Facilities leased to multiple tenant
companies for their IT
infrastructure, offering redundancy
and flexibility. Typical size: 5-20
MW.

The Campus
Hyperscale Cloud Data Centers
Large-scale facilities built and
operated by a single organization for
AI model training and cloud
computing platforms. Typical size:
20-100+ MW.

Attachment Staff Report Item 12A
Managing the Large -Load Development Pipeline
Ava does not control land use approvals or electric interconnections, but it has a strong interest in early visibility into potential large-load projects to align planning assumptions with real-world outcomes. 1 Direct Developer Engagement
Engage early with developers to understand
project maturity, timing, and likelihood of
completion.
2 Coordination with Local Planning
Collaborate with city and county staff who
have insights into zoning, permitting, and
development priorities.
3 Ongoing Information Sharing with PG&E
Maintain communication regarding
interconnection activity, infrastructure
constraints, and energization timelines.

This multi-pronged approach ensures Ava's load forecasts and procurement planning are based on credible project information, reducing the risk of misaligned energy procurement or rate design.

Attachment Staff Report Item 12A
Ava’s Large Load Engagement Strategy
Ava: Aligning Planning Outcomes
Attachment Staff Report Item 12A
Benefits for Ava and Its Member Communities
When managed with appropriate safeguards, large electric loads present strategic opportunities for Ava and its member communities.

Economic Benefits
1 gigawatt (GW) of new data center load could
support:
▪5,000 construction jobs
▪500 permanent technology jobs
▪more than 11,000 construction-related support jobs
▪nearly 3,000 permanent associated jobs
▪ $125–$175 million in incremental property tax revenue ▪$250–$300 million in additional sales tax revenues.

Source: PG&E
Community Advantages
• Spreads fixed administrative and program costs over a larger sales base.
• Supports long-term clean energy procurement at greater scale.
• Enables investments in new clean resources and grid
infrastructure.
• Provides direct and ongoing revenue streams to host
cities through Local Utility Users Tax (UUT).

Ava aims to protect existing customers while responsibly managing growth, ensuring costs and risks are borne by the customers driving the load.

Attachment Staff Report Item 12A
Looking Ahead
• Future Information Updates
• Discussions on large load tariffs or non-standard rate agreements
Any future proposals would be brought forward separately for Board consideration, with opportunities for policy
discussion and public input.
15 Attachment Staff Report Item 12A
Attachment Staff Report Item 12A
Item 6: Resolution Adopting the California
Energy Demand Forecast, 2025-2045
January 21, 2026 Business Meeting
Nick Fugate, Lead Forecaster
Energy Assessments Division, Demand Analysis Office
Attachment Staff Report Item 12B
Why does CEC forecast energy demand?
Warren-Alquist Act
Established the CEC
Public Resources Code
25301(a)
Requires the CEC to "conduct
assessments and forecasts of all
aspects of energy industry supply,
production, transportation,
delivery and distribution, demand,
and prices.”
Source: CEC
Source: CEC 2
Attachment Staff Report Item 12B
Benefits & Purpose of the Forecast
 Foundational for procurement and system planning in the state  Informs:
Resource adequacy requirements
Integrated resource planning
Reliability assessments
Transmission system planning
Distribution system planning
Attachment Staff Report Item 12B
Updates to the Forecast
 Minimal modeling changes to preserve stability
 Refreshed economic/demographic projections
 Updated inputs and scenario design for:
PV and battery storage
Data centers
Energy efficiency
Fuel substitution
Transportation electrification
 New analysis of utility “known loads”
Attachment Staff Report Item 12B
Assessing Known Loads
What are Known Loads?
 Energization requests at the distribution system level  Project-level data from each Investor-owned utility
Potential Reliability Risks
 Gap between transmission and
distribution planning
 Backlog of projects that exceed
growth in the IEPR forecast
Other Considerations
 First year that CEC has used the
known load data
 Lack of historical record
 Interaction with CEC modeling
 Balance resource procurement
obligations with grid planning
5 Attachment Staff Report Item 12B
Joint Agency and CAISO Decision on
Forecast Composition by Use Case
System/Flex
Resource
Adequacy
Local Capacity
Studies
Transmission Local
Studies
Integrated Resource Planning
& Transmission Bulk System
Forecast
Composition
2025 IEPR Planning
without Known
Loads
2025 IEPR Local
Reliability without
Known Loads
2025 IEPR Local
Reliability Forecast
with Known Loads
2024 IEPR Planning Forecast
 Monitor known loads
throughout 2026 for 2027
reliability
 Run CEC’s summer
reliability assessment with
and without known loads
 CAISO to use separate
scenarios for Local Capacity
Studies and Transmission
Local Studies
 Uncertainty with
known loads, data
centers
 Stability in resource
and infrastructure
planning
6 Attachment Staff Report Item 12B
Statewide Electricity Sales Forecast
Source: CEC
Attachment Staff Report Item 12B
Statewide Gas Sales Forecast
Source: CEC
Gas sales forecast implied by electricity system planning assumptions Attachment Staff Report Item 12B
CAISO System Annual Peak Forecast
Source: CEC 9
Attachment Staff Report Item 12B
Main Drivers of Peak Load Growth
Source: CEC 10
Attachment Staff Report Item 12B
2025 IEPR Forecast Workshops and Meetings
Meeting Type Topic Date
IEPR Workshop California’s Economic Outlook February 26 DAWG Meeting Economic and Demographic Inputs and Data
Center Forecasting
July 16
IEPR Workshop Inputs and Assumptions August 6
DAWG Meeting Load Modifier Assumptions August 18
IEPR Workshop Load Modifier Scenario Updates August 26
DAWG Meeting Draft Load Modifier Results October 30
IEPR Workshop Draft Load Modifier Results November 13
IEPR Workshop Draft Overall Forecast Results December 17 DAWG Meeting Revised Results & Forecast Composition January 5, 2026 2025 IEPR Workshops & DAWG Meetings 11
Attachment Staff Report Item 12B
Staff Recommendation
 Adopt the California Energy Demand Forecast, 2025-2045  Approve staff’s recommendation that this action is exempt from CEQA 12 Attachment Staff Report Item 12B
1 of 18

COMMITTEES ON UTILITIES AND ENERGY

AND PRIVACY AND CONSUMER PROTECTION

COTTIE PETRIE-NORRIS AND

REBECCA BAUER-KAHAN
CHAIRS
Wednesday, January 28, 2026
1:30pm
1021 N Street, Room 1100

JOINT OVERSIGHT HEARING
AI’s Energy Impacts
Findings
• Artificial intelligence (AI)-driven data center growth is occurring at a scale and speed that outpaces most existing electricity planning frameworks. Forecasting methods, interconnection processes, and cost-allocation rules were not designed for large, uncertain, and highly concentrated loads arriving simultaneously across multiple regions.
• California has so far avoided rate impacts from data center growth, but that outcome is not guaranteed. High electricity prices, project-by-project review, and existing system capacity have limited impacts to date; however, growing transmission-level load heightens the risk of future cost shifts.
• New data center customers can reduce – or increase – rates depending on how infrastructure costs and revenue are allocated. Bill reductions only occur when new loads cover their full marginal and long-term costs; otherwise, ratepayers face higher bills, stranded assets, and “ghost load” risk.
• Uncertainty around data center demand materially increases financial and reliability risks for the grid. Overstated or duplicative load projections can drive unnecessary procurement and transmission investment, as evidenced in other regions, with costs borne by customers when projected load fails to materialize. • Current regulatory pathways for backup generation and flexible demand may conflict with California’s environmental and equity goals. Expanded reliance on diesel backup generation, especially under streamlined siting and flexibility programs, could increase localized pollution and undermine emissions objectives if not carefully constrained.

Attachment Staff Report Item 12C
2 of 18
The increasing computational requirements of AI have fueled a wave of data center development, with an estimated 50 gigawatts (GW) of new data center capacity added to the global development pipeline in 2024 alone.1 This is an unprecedented level of demand in an extraordinarily short amount of time. For context, 50 GW is roughly the entire electricity demand of California during its hottest, most stressed hours – essentially the load of the whole state during extreme heat events.
2 This rapid expansion of data centers is driving an unprecedented demand for energy. In 2024, data centers in the U.S. used approximately 200 terawatt-hours (TWh) of electricity, roughly what it takes to power Thailand for a year.3 As a result, access to electricity is a top concern shaping the development of digital infrastructure.

At the same time, access to affordable electricity is a top concern shaping the public discourse nationwide, particularly in areas with large data center concentrations. Across the country – from Georgia4to Texas,5 Ohio6to New York,7to the hardest hit area in Northern Virginia 8 – the
impact of AI-fueled data center growth has become a kitchen-table issue, replete with yard signs,9local moratoria,
10 and high electricity bills fueling a frustrated public.11,12 According to the U.S. Energy Information Administration, residential utility bills rose 5% nationally by October 2025;
13 however, this was much higher in data center-rich regions, with Virginia, Ohio, and Illinois experiencing double-digit increases.14
Yet these challenges have not yet surfaced in California. Despite being one of the top data center markets in the country,15 California has not seen an increase in electricity rates due to data center growth, according to the California Public Utilities Commission (CPUC).16 Although data 1 iMasons. State of the Digital Infrastructure Industry – Annual Report 2025. 2025. https://imasons.org/publications/state-of-the-digital-infrastructure-ind ustry-annual-report-2025/ 2 September 6, 2022 reached a peak load of approximately 52 GWs, the highest on record. California’s average annual demand is between 25-30 GWs.
3 James O’Donnell and Casey Crownhart, “We did the math on AI’s energy footprint. Here’s the story you haven’t heard.” MIT Technology Review, May 20, 2025; https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/ 4 Kala Hunter, “How Georgia became the ‘wild west’ of data centers. Is transparency on the horizon?” Georgia Public Broadcasting, September 10, 2025; https://www.gpb.org/news/2025/09/10/how-georgia-became-the-wild-west-of-data-centers-transparency-on-the-horizon 5 Jason Plautz, “Trump, atoms, AI and the Texas data center gusher,” Politico, January 4, 2026; https://www.politico.com/news/2025/12/23/fermi-america-data-center-amarillo-texas-00701800?_sp_pass_consent=true 6 Cliff Pinckard, “Ohio ranks fifth in the country for data centers. To power them, they’re going nuclear.” Cleveland.com (formerly The Plain Dealer), January 20, 2026. https://www.cleveland.com/metro/2026/01/ohio-ranks-fifth-in-the-country-for-data-centers-to-power-them-theyre-going-nuclear-the-wake-up-for-tuesday-jan-20-2026.html 7 Johan Sheridan, “Data center boom straining power grid as New York asks who should pay,” ABC News 10, October 24, 2025; https://www.news10.com/news/rising-energy-costs-new-york/ 8 Liam Bowman, “Concerns over data centers drive local election in Northern Virginia,” The Washington Post, November 9, 2025. https://www.washingtonpost.com/dc-md-va/2025/11/08/prince-william-county-gainseville-election/ 9 Ryan Murphy and Emily Feng, “Why more residents are saying ‘No’ to AI data centers in their backyard;” NPR News; July 17, 2025; https://www.npr.org/2025/07/17/nx-s1-5469933/virginia-data-centers-residents-saying-no 10 Anna Lynn Winfrey, “Another Columbus-area community passes 90-day moratorium on data centers,” Columbus Dispatch, December 15, 2025; https://www.dispatch.com/story/news/local/communities/dublin/2025/12/15/second-central-ohio-community-washington-twp-to-ban-data-centers-for-90-days-wants-dublin-to-join-in/87719185007/ 11 Marc Levy and Jesse Bedayn, “Voters’ anger at high electricity bills and data centers looms over 2026 midterms,” The Associated Press, November 8, 2025; https://apnews.com/article/2026-election-utility-bills-ai-data-centers-13703f61d1397612fd067e69b9093116 12 J. Saul, et al., “AI Data Centers are Sending Power Bills Soaring,” Bloomberg Technology, September 29, 2025; https://www.bloomberg.com/graphics/2025-ai-data-centers-electricity-prices/ 13 As compared to October 2024; EIA information released on December 23, 2025 with data for October 2025. https://www.eia.gov/electricity/monthly/update/
14 Kimball and Cortes, “Data centers are concentrated in these states. Here’s what’s happening to electricity prices.” CNBC, November 14, 2025; https://www.cnbc.com/2025/11/14/data-centers-are-concentrated-in-these-states-heres-whats-happening-to-electricity-prices-.html 15 See Figure 2 below
16 CPUC, “California Data Center Development & Energy Needs: FAQs,” 12.22.2025.

Attachment Staff Report Item 12C
3 of 18
centers use significant amounts of electricity, they still represent a small share of California’s overall demand and are typically sited and managed in ways that have limited their impact on statewide electricity rates. In investor-owned utility (IOU) territories, the CPUC notes that commercial and industrial (C&I) rate structures often require project-by-project review, which helps identify data center–related system upgrades and assign those costs to the data center driving them. In addition, California’s above-average electricity prices can discourage electricity-hungry data center development, except where proximity to users is required to meet latency needs.

These trends may be changing. For instance, Pacific Gas & Electric (PG&E) residential electricity rates were 11% lower by January 2026 than in January 2024,17 in contrast to other California IOUs. But PG&E also reported a rise in future data center development during this time, from 5.5 GWs in February 202518 to 10 GWs by July 2025.19 It will take years for this data center load to come online. But as that happens, the generation and new transmission lines to serve that load will also come online – the cost of which is traditionally shared amongst all customers.

The emergence of extensive data center electricity load, driven by the growth in AI computing, marks a rare opportunity in the energy industry to reassess existing rate design and customer protections. The promise of data center energy consumption is alluring: that increased electricity sales from data centers will cover all the new costs to serve those data centers and may even offset existing system costs. Such is the story for other large loads, like transportation electrification (TE), with the California Public Advocates Office (PAO) finding TE “may cause downward pressure on electric rates.”20 The timing can feel almost too perfect: as California undertakes systemwide decarbonization and infrastructure renewal, a new, well-funded customer emerges, apparently willing21 to shoulder the cost of long-deferred investments. However, as the PAO cautions for TE, and as applies equally to data-center development, “to achieve this downward pressure on rates, effective management of multiple factors will be required, including efficient infrastructure buildout and cost constraints.”22 California is better situated than many other parts of the country to accommodate data center driven load growth. California’s early adoption of resource planning, for both decarbonization and reliability, have provided a cushion, such that the California grid has significant capacity available in periods of low customer demand.23 California also boasts a legacy of innovation, 17 PG&E press release, “PG&E to Lower Electric Prices on Jan. 1, Fourth Decrease in Two Years,” December 30, 2025; https://investor.pgecorp.com/news-events/press-releases/press-release-details/2025/PGE-to-Lower-Electric-Prices-on-Jan--1- Fourth-Decrease-in-Two-Years/default.aspx
18 PG&E press release, “PG&E Accelerating Connection of New Data Centers throughout Northern and Central California,” February 13,2025; https://investor.pgecorp.com/news-events/press-releases/press-release-details/2025/PGE-Accelerating-Connection-of-New-Data-Centers-throughout-Northern-and-Central-California/default.aspx 19 PG&E press release, “PG&E Data Center Demand Pipeline Swells to 10 Gigawatts with Potential to Unlock Billions in Benefits for California,” July 31, 2025; https://investor.pgecorp.com/news-events/press-releases/press-release-details/2025/PGE-Data-Center-Demand-Pipeline-Swells-to-10-Gigawatts-with-Potential-to-Unlock-Billions-in-Benefits-for-California/default.aspx 20 Pg. 14, Public Advocates Office, Distribution Grid Electrification Model 2025 – Study and Report, October 2025. https://www.publicadvocates.cpuc.ca.gov/-/media/cal-advocates-website/files/press-room/reports-and-analyses/251030-public-advocates-office-distribution-grid-electrification-model-2025.pdf 21 Amrith Ramkumar, “Microsoft Makes New Data-Center Pledges After Local Backlash,” The Wall Street Journal, January 13, 2025.
22 Pg. 14, PAO October 2025, Ibid.
23 The transmission system uses less than 40% of its capacity on average. Pg. 8, L. Min, et al., Powering AI at Speed in California, Stanford Precourt Institute for Energy, September 2025; https://drive.google.com/file/d/14NhXPKynQCjbBC99xVk6tA5pqCD9U052/view Attachment Staff Report Item 12C
4 of 18
from early data center adoption in Santa Clara to energy efficiency requirements24 to historic approaches to ensure adequate capacity.
25 The pace of AI deployment has strained most aspects
of current energy governance and cost allocation; yet California’s high electricity costs may have provided the rare benefit of slowing recent in-state development, enabling California regulators the time to apply effective safeguards and lessons learned from other states.

The purpose of this hearing is to discuss the impact of AI and corresponding data center growth on the energy sector, both nationwide and in California. The hearing will detail impacts specific to California’s electricity system – from maintaining renewables goals, to ensuring system reliability, to protecting affordability. While data center electricity and water consumption are related, sometimes inversely,26 this hearing will focus just on the electric side. Other committees may hold hearings later this session focused on other resource constraints, including water.

Panelists will be asked to speak to needed safeguards to ensure data center development in California does not lead to runaway energy costs, stranded assets, or negative environmental impacts.

I. The Data Center Ecosystem.

Data centers pose an emerging challenge for California’s energy system. But not all data centers are built the same. Their power demands, latency needs, and location requirements vary significantly depending on their function.

Some key players include:
• Enterprise Data Centers – Private data centers that serve proprietary digital operations (e.g., banks, healthcare organizations, governments). The homeowners of the data center world.
• Colocation Data Centers – Facilities where multiple tenants rent space to host their own servers and technology equipment, without building or managing their own space. These are the landlords of the data center world.
• Hyperscale Cloud Providers – Large-scale operators that deliver rapid computing power and cloud-based storage to support AI and other data-intensive applications. Examples include Amazon Web Services (AWS), Microsoft Azure, and Google Cloud. These are like colleges or hospitals, where the entire campus is owned and managed by a single entity; they sell services rather than space.
• Edge Data Centers – Smaller, distributed facilities located closer to end-users or data sources. They support processes that rely on low latency. They are critical for services like autonomous vehicles and content delivery, often operating in conjunction with larger cloud or colocation providers to offload demand from centralized data centers. These are like the neighborhood corner store – built for speed and convenience, not scale. 24 CEC, “Computer Rooms & Data Centers Fact Sheet,” August 2023. https://www.energy.ca.gov/sites/default/files/2023- 09/2022_CEC-Computer_Room_and_Data_Centers_ADA.pdf
25 Namely our Resource Adequacy program; https://www.cpuc.ca.gov/industries-and-topics/electrical-energy/electric-power-procurement/resource-adequacy-homepage 26 For data center cooling, specifically, the less water you use (water towers), the more electricity you have to use (air chillers), and vice versa.

Attachment Staff Report Item 12C
5 of 18
Data centers rely on long-term contracts and make money by leasing space and selling computing and storage services. Because building data centers requires very large upfront investments – often hundreds of millions or even billions of dollars – strong financing is necessary. Their profitability depends on factors such as energy costs, customer demand, lease terms, and how efficiently the facilities are operated.

The emergence of AI has led to a shift in data center architecture, from traditional centralized computing toward more decentralized networks. Traditionally, markets such as Loudoun County, Virginia and Silicon Valley, California have been key data center hubs due to their infrastructure and connectivity, as shown by the dark blue bars in Figure 1.27 As demand for real-time processing increases, more decentralized approaches – such as edge computing – are changing where data centers are built. These trends are expanding data center development beyond traditional metro areas into regions with more favorable energy costs, permitting conditions, and transmission access, as shown by lighter blue bars in Figure 1. These trends have influenced the scale and location of data center development in California.

Figure 1. Top 10 U.S. data center markets by future power capacity (including current inventory and under construction & planned).

In California, data center development has been concentrated in Santa Clara and Los Angeles, due to proximity to fast fiber-optic connections and lower electricity rates served by those cities’ municipal utilities. More recently, Pacific Gas & Electric has received a surge of requests for service to new data centers. Initially those requests were in and around the South Bay; however, the utility reports expanding development of large load customers in other regions of its service territory.28
II. The Dawn of Data.
27 L. Min, W. Chueh, and I. Ehrenpreis, Powering AI at Speed in California, Precourt Institute for Energy [Stanford University] (Sep. 2025), p. 7. https://drive.google.com/file/d/14NhXPKynQCjbBC99xVk6tA5pqCD9U052/view 28 Michael Medieros, “PG&E Written Testimony;” Letter to the Little Hoover Commission; December 8, 2025. https://lhc.ca.gov/wp-content/uploads/PGE-LHC-Written-Testimony.pdf Attachment Staff Report Item 12C
6 of 18
In 2024, global data generation surpassed 149 zettabytes (trillions of gigabytes).29 As put into perspective by researchers at University of Texas at Austin: “if each gigabyte were a single-page document, the resulting stack of [these] papers could reach the Moon and back ∼20 times.”30 This scale of data reflects the rapid growth of connected devices, AI-driven queries, and cloud-based services that depend on data centers for real-time processing and storage.

It is the growth in AI-computing that has
driven most of the growth in data center
development and construction. However, the
data and energy needs vary depending on the
type of AI-computing. Traditionally, there are
two types of AI workloads: training and
inference. Training is when AI learns from
large datasets; whereas inference is when that
trained model is used by customers.

Data Hogs are Energy Hogs. As noted in
reports by the Assembly Committee on
Privacy and Consumer Protection,31 staggering
quantities of data are required to train the most
advanced AI models. For example, GPT- 4 –
the large language model (LLM) embedded in
ChatGPT 4 – is reported to have been trained
on roughly 10 trillion words of text.32,33
Adjusting the model’s 1.8 trillion parameters
continuously as it was exposed to this vast
corpus required trillions upon trillions of
computations, which were performed by
running approximately 25,000 expensive,
energy-consuming microchips for nearly 100
days nonstop, at an estimated cost of $63
million.34 It is estimated this training
consumed 50 gigawatt-hours (GWhs) of
29 D. Ewim et al. “Impact of Data Centers on Climate Change: A Review of Energy Efficient Strategies”. The Journal of Engineering and Exact Sciences 9.6 (2023), 16397–01e. https://www.researchgate.net/figure/Global-annual-GHG-emissions-from-major-industries-Source-Kilgore-2023_fig1_373295068 30N. Ling, et. al, Data Center Growth in Texas: Energy, Infrastructure, and Policy Pathways; Bureau of Economic Geology, The University of Texas at Austin; December 2025; https://www.beg.utexas.edu/files/cee/Data_Center_White_Paper_BEG.pdf 31 Most recently in their background paper for their December 8, 2025, joint hearing on “Artificial Intelligence and Copyright” https://apcp.assembly.ca.gov/system/files/2025-12/background-paper-dec-8-stanford-ai-and-copyright-joint-info-hearing_updated.pdf 32 Schreiner, “GPT-4 architecture, datasets, costs and more leaked,” The Decoder (Jul. 11, 2023), https://the-decoder.com/gpt-4- architecture-datasets-costs-and-more-leaked/
33 Begum, “OpenAI Releases GPT-4: A Smarter and Faster AI-Language Model with ‘Human-level Performance,’” Vocal Media (2023), https://vocal.media/01/open-ai-releases-gpt-4-a-smarter-and-faster-ai-language-model-with-human-level- performance. 34 Ludvigsen, “The carbon footprint of GPT-4,” Medium (Jul. 18, 2023), available at https://medium.com/data- science/the-carbon-footprint-of-gpt-4-d6c676eb21ae.

Box 1: A Note on Data Availability
Details about AI’s current and future energy
impact are fairly murky. As noted by researchers
at Lawrence Berkeley National Lab, “the lack of
direct energy data available in a sector with
rapidly evolving technologies limits this
analysis…” a
The researchers go on to note, “very few
companies report actual data center electricity
use and virtually none report it in context of IT
characteristics…” b
As such, values reported on AI or data center
energy use in this document are estimates with
large uncertainty.

Such vagueness may be fine for purposes of this
discussion, but as highlighted below, can greatly
limit state energy planners’ abilities to
accurately forecast grid system needs. This
dearth of information motivated both AB 222
(Bauer-Kahan, 2025) and the recently
introduced AB 1577 (Bauer-Kahan, 2026).
a: pg. 7, Shehabi, Arman, et al. "2024 United States
Data Center Energy Usage Report." December 2024,
Lawrence Berkeley National Laboratory, Berkeley,
California. LBNL-2001637; https://eta-publications.lbl.gov/sites/default/files/2024-12/lbnl-2024-united-states-data-center-energy-usage-report_1.pdf b: pg. 68, Shehabi, Ibid.

Attachment Staff Report Item 12C
7 of 18
energy, enough to power San Francisco continuously for three days.35 However, as noted in MIT Technology Review, “inference, not training, represents an increasing majority of AI’s energy demands. …It is now estimated that 80-90% of computing power for AI is used for inference.”36 While rough estimates exist for how much energy each AI-query consumes – famously, OpenAI CEO Sam Altman has stated each ChatGPT query consumes approximately 0.34 watt-hours of electricity, enough to power a lightbulb for a few minutes37– in reality, the type and size of the model and your selected output (language, image, video) can make one query thousands of times more energy-intensive and emissions-producing than another. One billion of these inquiries every day for a year – well below the self-reported values of OpenAI38– would mean over 109 GWhs of electricity, enough to power 10,400 U.S. homes for a year, according to researchers writing in MIT Technology Review.39 The type of AI workloads will also strongly shape where future data centers are built. Training large AI models requires enormous computing power but is relatively insensitive to latency, allowing these facilities to be sited in remote areas with access to abundant, low-cost electricity.

In contrast, inference workloads support real-time applications and are highly latency-sensitive, which favors data centers located close to end users. As a result, energy-intensive training facilities are increasingly being built in rural or industrial regions, while smaller edge and micro data centers are expanding in urban areas to support inference-driven services such as autonomous driving and immersive media.

Exponential growth doesn’t feel
exponential at first. As shown in
Figure 2, U.S. data center annual
energy use prior to 2016 was
relatively stable at about 60 TWh.

Despite the construction of new data
centers to serve the rise of cloud-based online services, such as Netflix, increases in efficiency kept
this growth consumption relatively
flat. However, starting in 2017,
electricity consumption accelerated
nationwide primarily due to AI
computing. By 2023, data centers
were 4.4% of total U.S energy
consumption. By 2028, data centers
are predicted to use between 6.7%
and 12% of the country’s
35 James O’Donnell and Casey Crownhart, “We did the math on AI’s energy footprint. Here’s the story you haven’t heard.” MIT Technology Review, May 20, 2025; https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/ 36 O’Donnell and Crownhart, Ibid.
37 Kwan Wei Kevin Tan, “Sam Altman says the energy needed for an average ChatGPT query can power a lightbulb for a few minutes,” Business Insider, June 10, 2025; https://www.businessinsider.com/how-much-energy-does-chatgpt-use-average-query-watts-altman-2025-6 38 Which in July 2025 estimated 2.5 billion prompts every day; Emma Roth, “OpenAI says ChatGPT users send over 2.5 billion prompts every day,” The Verge, July 21, 2025; https://www.theverge.com/news/710867/openai-chatgpt-daily-prompts-2-billion 39 O’Donnell and Crownhart, Ibid.

Figure 2. Total U.S. data center electricity use from
2014-2028.10
Attachment Staff Report Item 12C
8 of 18
electricity.40 At that point, AI alone could consume as much electricity annually as 22% of all U.S. households.41
Just how utilities, regulators, and tech companies will meet this mawing demand is uncertain.

New resources must be procured, and new
transmission is likely needed to serve this
load, all at cost to either ratepayers or the
tech companies themselves. Yet given the
large uncertainty projected (grey, shaded
area in Figure 2), and the lack of transparent
data on current and future data center energy
needs (See Box 1), projections of future
energy demands or estimates of resultant
emissions are simply inadequate or
inaccurate. This forces energy planners to
assemble a puzzle with countless missing
pieces.

III. Predicting the Unpredictable.

Last Wednesday, January 21, 2026, the
California Energy Commission (CEC)
adopted their energy demand forecast for
2025-2045.42 This forecast is foundational
for resource procurement and system
planning in the state, as the output of the
forecast feeds into sequential planning
streams at the CPUC and California
Independent System Operator (CAISO),43 as
detailed in Box 2. In other words, the
forecast is critical in determining how much
energy generation to require utilities to
purchase, where to upgrade power lines, and
how to prevent future blackouts.

The 2025 IEPR provides electricity and gas
demand forecasts which reflect expected
impacts from economic projections,
including data center growth, electric vehicle
adoption, and other inputs.44 The final 2025
adopted plan anticipates almost 20 GWs of
load growth over the next 20 years in
40 Shehabi, Arman, et al. "2024 United States Data Center Energy Usage Report." December 2024, Lawrence Berkeley National Laboratory, Berkeley, California. LBNL-2001637; https://eta-publications.lbl.gov/sites/default/files/2024-12/lbnl-2024-united-states-data-center-energy-usage-report_1.pdf 41 O’Donnell and Crownhart, Ibid.
42 Item 6, CEC Business Meeting Agenda, January 21, 2026; https://efiling.energy.ca.gov/GetDocument.aspx?tn=268217&DocumentContentId=105381 43 a non-profit public benefit corporation regulated by the Federal Energy Regulatory Commission (FERC) who maintains operational control of ~80% of the state’s transmission grid. 44 CEC, “Resolution of the CEC Adopting the California Energy Demand Forecast, 2025-2045,” https://www.energy.ca.gov/filebrowser/download/9208?fid=9208 Box 2: A Primer on CA’s System Planning
California energy planning is conducted via
layered programs administered by several state
entities, chiefly: the CEC, the CPUC, and
CAISO. These entities have a memorandum of
understanding to provide clarity and specificity
in how information and data are transmitted
and utilized by each entity. The main planning
processes at each organization include:
• The Integrated Energy Policy Report
(IEPR) – CEC – every 2 years: forecasts all
aspects of energy supply and demand. The
demand forecast from the IEPR is a
primary input for planning at the CPUC
and CAISO.
• Integrated Resource Plans (IRP) – CPUC –
every 2 years: process to ensure long-term
resource procurement. Energy purchases
are still conducted by individual energy
suppliers, but the IRP ensures adequate
resources are bought to meet the IEPR
forecast and electricity sector greenhouse
gas goals. Serves as a primary input for
planning at the CAISO.
• The Transmission Planning Process (TPP)
– CAISO – annually: estimates the
transmission infrastructure needed to match
the IRP portfolio provided annually by the
CPUC with the demand forecast provided
by the CEC. Once the final TPP is
approved by the CAISO Board, the
resulting approved infrastructure begins the
development process, including permitting,
licensing, and competitive solicitations as
applicable.

Attachment Staff Report Item 12C
9 of 18
CAISO’s area, with close to 5 GWs arising from data centers. While significant, the adopted forecast shows a downward adjustment of almost 5 GWs by 2040 between the 202445 and 2025 IEPRs, largely due to the treatment of data centers’ “known loads” in the 2025 models. If “known loads” are included, the peak forecast swells by almost 10 additional gigawatts by 2045 – a significant range of uncertainty.

All Models are Wrong, But Some are Useful.46 As reported in Politico, Nick Fugate, the CEC’s lead forecaster, has noted that uncertainty “has increased significantly” in the last few IEPRs.47 “Known loads” are new inputs into the 2025 IEPR modeling, and are based on IOU-submitted project-level information, including energization requests and expected in-service dates. As shown in Figure 3, significant uncertainty exists around these known load requests.48 Because these known loads lack a historical record within the CEC, the final 2025 forecast “Set Agreement” between the CEC, CPUC, and CAISO declines to advance the forecast with the known load values. Instead, the “Set Agreement” notes the CEC staff will review historical known loads throughout 2026 to “confirm assumptions informed by the IOUs” and “will continue to monitor data center applications for energization…”49 Figure 3. Capacity Requests, as reported to the CEC from each utility as of August 2025.48 (Key:

PG&E = Pacific Gas & Electric; SVP = Silicon Valley Power; SCE = Southern California Edison; SDG&E = San Diego Gas & Electric)
Interestingly, the adopted 2025 agreement permits the use of known loads to inform local transmission planning at the CAISO but not reliability planning at the CPUC.50 Parties representing large energy users, including data centers, have raised concern that this bifurcated inclusion of known loads will likely be “insufficient to serve projects.”51 However, the “Set 45 ~ 67 GWs anticipated by 2040 in the 2024 IEPR (pg. 38, https://efiling.energy.ca.gov/GetDocument.aspx?tn=266141) versus ~62 in the 2025 IEPR (slide 9, https://www.energy.ca.gov/filebrowser/download/9328?fid=9328) 46 Quote largely attributed to statistician George Box. 47 Noah Baustin, “California’s new grid hog isn’t who you think,” Politico, January 21, 2026; https://www.politico.com/newsletters/california-climate/2026/01/21/californias-new-grid-hog-isnt-who-you-think-00740515 48 Slide 5, presentation by CEC Manager of Demand Analysis Heidi Javanbakht to the Little Hoover Commission, December 11, 2025; https://lhc.ca.gov/wp-content/uploads/5-Heidi-Javanbakht-CEC.pdf 49 Pg. 6, CEC, “Single Forecast Set Agreement,” 2025 IEPR Forecast Supporting Documentation, filed January 23, 2026; https://efiling.energy.ca.gov/GetDocument.aspx?tn=268288&DocumentContentId=105461 50 Pg. 7, CEC Set Agreement, Ibid.
51 Meredith Alexander, “Comments on the 2025 IEPR Draft Electricity Demand Forecast,” January 9, 2026; https://efiling.energy.ca.gov/GetDocument.aspx?tn=268213&DocumentContentId=105376 Attachment Staff Report Item 12C
10 of 18
Agreement” notes the CEC will closely monitor known load energizations throughout 2026, with the possibility of intervention in advance of 2027 should load growth outpace current forecast expectations.

While the demand forecast is independent of system cost, significant financial consequences do exist if the models prove grossly inaccurate. For instance, if the known loads were included in the planning forecast, at the current CPUC resource adequacy benchmark of $11.53/kilowatt-month, 52 the cost impact to serve that known load is in the range of $500 million for 2026- 2027.
53 This is a significant cost impact, especially given the uncertainty of whether this load will even materialize to pay for these costs.

The Ghost in the Model. Within the PJM Interconnection54– the largest power grid operator in the U.S., serving approximately 67 million customers from Chicago to New Jersey, including the “data center alley” of Northern Virginia – forecasted load growth has emerged as a key driver in utility cost surges over the last year. According to PJM’s Independent Market Monitor, load growth from new data centers was responsible for roughly $9.3 billion of the $14 billion regional capacity market bill for 2025-26, with costs escalating further over the next 2 years. PJM is also projecting an additional $10+ billion in transmission expenditures, largely caused by data centers.55
Yet many stakeholders are raising concern that these projections are inflated, with PJM recently adopting stricter vetting for its data center load forecast in response. 56 When projected energy
demand doesn’t materialize, it is colloquially termed “ghost load.” Grid planners may see “ghost load” for several reasons. Data center developers often forum shop across multiple states and utility territories at the same time, looking for the best tax incentives, access to land, quick electrical hookups, and favorable regulatory conditions. Because utilities plan for and report this potential load individually, the same data center can be counted multiple times across different jurisdictions – even if it is never ultimately built. Utilities and transmission operators then make investment decisions based on these projections and report them to their grid planners. Those planners, in turn, use the inflated estimates to procure capacity and plan new transmission, which can drive higher electricity costs for customers.

In response to the emergence of “ghost loads,” various states have adopted strategies to make developers provide more upfront investments before utilities spend any ratepayer money. In Illinois, developers are required to provide a $1 million deposit.57 In Indiana and Ohio, long-term take-or-pay contracts are required where developers pay for at least 85% of the transmission service they request for at least 10 years, even if their actual usage is less.58 And recently FERC ordered the PJM Interconnection to update its tariff to provide for interconnection of customers serving co-located load (i.e., data centers with onsite generation).59 This FERC order largely 52 Table 2, CPUC, “Market Price Benchmark Calculations 2025,” October 1, 2025; https://www.cpuc.ca.gov/-/media/cpuc-website/divisions/energy-division/documents/community-choice-aggregation-and-direct-access/2025-mpbs.pdf 53 using an ~18% planning reserve margin, estimating ~$465-$565 million in additional cost 54 “PJM” comes from the original name: the “Pennsylvania-New Jersey-Maryland Interconnection” 55 IMM for PJM, “Analysis of the 2027/2028 RPM Base Residual Auction Part A,” January 5, 2026. 56 Ethan Howland, ”PJM trims near-term load forecast on stricter data center vetting, economic outlook,” Utility Dive, January 15, 2026; https://www.utilitydive.com/news/pjm-interconnection-load-forecast-data-centers/809717/ 57 John Pletz, “AI gold rush fuels ComEd crackdown on data center speculators,” Chicago Business, June 24, 2025; https://www.chicagobusiness.com/utilities/comed-seeks-bigger-deposits-ai-era-data-centers 58 Nick Evans, “Ohio Manufacturers’ Association challenges new utility billing for data centers,” Ohio Capital Journal, November 13, 2025; https://ohiocapitaljournal.com/2025/11/13/ohio-manufacturers-association-challenges-new-utility-billing-for-data-centers/ 59 FERC Order to PJM, December 18, 2025, Docket EL25-49-000; https://www.ferc.gov/media/e-1-el25-49-000-0# Attachment Staff Report Item 12C
11 of 18
paves the way for the “Bring Your Own Generation” pathway under consideration at the PJM Board.60
Implementing these policies at the state level has been shown to lead to immediate reduction in ghost load within a given state. In Ohio, for instance, AEP Ohio saw a 50% decrease in projected load upon implementing its large load tariff.61 In California, PG&E’s Rule 30 tariff application includes some of these reforms – such as minimum demand charges, early termination fees, and paying upfront some of the costs to interconnect.62 However, the PAO and others have recommended additional modifications to Rule 30 to ensure maximum ratepayer protections.63 IV. Narrow Tolerances, Abrupt Drops.

On July 10, 2024, a device that protects electrical systems from damage during lightning strikes – a “lightning arrestor” – failed, resulting in a fault on a 230-kilovolt transmission line in the Eastern Interconnection. The transmission line was set up to automatically attempt restarting three times from each end of the line after a fault. Because both ends of the line were doing this, the line repeatedly shut off and restarted, causing six brief faults in just over a minute. The protection system worked as designed by detecting each fault and eventually (in 82 seconds) safely shut off the line. Nothing about the behavior of the transmission line’s protective equipment was unusual.

What was unusual was at the same time and near the same area, approximately 1.5 GWs of load dropped off the grid. As reported by the North American Electric Reliability Corporation’s (NERC) investigation of the event, none of the load was disconnected from the system by utility equipment; rather, the load was disconnected by the customer, specifically data center customers.
64 NERC concluded that a protective/control scheme on the data centers’ uninterruptible power supplies (UPS) – which are generally either power electronics that switch load to a battery bank or a flywheel that switches load to a diesel engine – was set to a particularly sensitive setting, causing over a gigawatt of load to drop off the grid and remain off for hours.

While grid operators had to act to stabilize the system following this load drop, the disturbance did not result in significant operational issues. However, NERC notes similar incidents have occurred in other interconnections with both cryptocurrency and oil and gas loads.

While CAISO, and many other grid operators, routinely plan for the sudden loss of large electric supply,
65 little planning has been done nationwide to manage the sudden loss of such significant load.66 NERC recommends in its Incident Report that transmission operators enter into 60 Joint Governor’s Joint Proposal for CIFP LLA, November 5, 2025; https://www.pjm.com/-/media/DotCom/committees-groups/cifp-lla/2025/20251106/20251106-item-04h---joint-dcc-and-governors-proposal.pdf 61 Zachary Jarrell, “AEP Ohio cuts its data center demand forecast in half,” Biz Journal, September 25, 2025. https://www.bizjournals.com/columbus/news/2025/09/25/aep-ohio-data-center-demand-forecast.html 62 A. 24-11-007; “Application of PG&E for Approval of Electric Rule No. 30 for Transmission-level Retail Electric Service,” filed November 21, 2024. https://docs.cpuc.ca.gov/PublishedDocs/SupDoc/A2411007/7851/547330908.pdf 63 Akiya and McCormack,” Public Advocates Office Opening Brief,” A. 24-11-007, October 24, 2025; https://docs.cpuc.ca.gov/PublishedDocs/Efile/G000/M584/K972/584972803.PDF 64 Pg. 2, NERC, “Incident Review: Considering Simultaneous Voltage-Sensitive Load Reductions;” January 8, 2025; https://www.nerc.com/globalassets/our-work/reports/event-reports/incident_review_large_load_loss.pdf 65 For CAISO, this single contingency is the loss of the Diablo Canyon nuclear power plant. WECC rules require 6% contingency reserves for this loss. CAISO also requires electrical entities to maintain an additional 9% in reserves to account for other potential plant outages or higher-than-average peak demand, leading to a total 15% planning reserve margin. 66 CAISO, “California ISO Planning Standards,” February 2, 2023; https://www.caiso.com/Documents/ISO-Planning-Standards-Effective-Feb22023.pdf Attachment Staff Report Item 12C
12 of 18
agreements with large load customers to ensure better coordination when reconnecting their systems. CAISO is already exploring various “ride-through” characteristics of various customer types and evaluating the potential impacts of on-site generation supporting these large loads on the transmission grid.67
V. Sustainability Goals and Procurement Options.

Many hyperscalers have continued to uphold corporate sustainability commitments even as AI-related computing demand has surged, channeling investment not only into nuclear power but also emerging clean energy technologies and carbon capture. 68 Big tech firms seem to be single-handedly reviving the U.S. nuclear energy market, with nuclear project announcements ranging from Meta69 to AWS70 to Microsoft’s famous restart of Three Mile Island.71 As data center electricity demand grows, these commitments could help support further investment in underdeveloped carbon-free resources, though translating this potential into durable systemwide benefits is not guaranteed.

In 2024, Microsoft touted an agreement for 10.5 GWs of new renewable capacity in the U.S. and Europe, a scale Microsoft reported as “almost eight times larger than the largest corporate power purchase agreement (PPA) ever signed.”72 While this generation does not directly power its operations, Microsoft also reported an agreement with Powerex to match hourly demand at a Washington datacenter with direct deliveries of carbon-free power on a 24-hour basis throughout the year.73
Microsoft is hardly alone in these sustainability achievements. Amazon has committed to matching all the electricity consumed by its operations with 100% renewable energy, a goal it achieved 7 years early in 2023. Amazon achieved this by investing billions in global solar and wind projects, as well as providing on-site solar directly on Amazon buildings.74 While commendable, this investment does not mean Amazon’s operations are directly powered by renewable energy. Instead, an equivalent amount of renewable electricity is delivered to the grid and credited through accounting mechanisms.

Even with this approach, Amazon’s sustainability report shows that emissions from its direct operations increased by 7% in 2023, and its overall carbon footprint has grown by 34% since 2019.75 Recent reporting from S&P Global noted “carbon-emitting sources have supplied 20%- 67 Pgs. 4-5, CPUC, “California Data Center Development & Energy Needs: FAQs,” 12.22.2025. 68 Spencer Kimball, “Exxon in advanced talks to power AI data centers with natural gas and carbon capture,” CNBC, October 31, 2025; https://www.cnbc.com/2025/10/31/exxon-ai-data-center-natural-gas-carbon-capture.html?msockid=3f21a963703c6fbc1a35bfda71ff6eac 69 Meta newsroom, “Meta Announces Nuclear Energy Projects, Unlocking Up to 6.6 GW to Power American Leadership in AI Innovation,” January 9, 2026; https://about.fb.com/news/2026/01/meta-nuclear-energy-projects-power-american-ai-leadership/ 70 US EIA, “Data center owners turn to nuclear as potential electricity source,” Today in Energy, October 1, 2024; https://www.eia.gov/todayinenergy/detail.php?id=63304
71 Casey Crownhart, “Why Microsoft made a deal to help restart Three Misle Island,” MIT Technology Review, September 26, 2024; https://www.technologyreview.com/2024/09/26/1104516/three-mile-island-microsoft/ 72 Bobby Hollis, “Accelerating the addition of carbon-free energy: An update on progress,” Microsoft blog, September 20, 2024; https://www.microsoft.com/en-us/microsoft-cloud/blog/2024/09/20/accelerating-the-addition-of-carbon-free-energy-an-update-on-progress/ 73 Powerex press release, “Powerex announces agreement to provide 24x7 hourly matching of carbon-free energy to Microsoft,” June 2023; https://powerex.com/sites/default/files/2023- 06/Powerex%20Announces%20Agreement%20with%20Microsoft%20for%2024x7%20Carbon-Free%20Energy.pdf 74 Amazon, “Amazon meets 100% renewable energy goal 7 years early;” Amazon News, last updated August 14, 2025; https://www.aboutamazon.com/news/sustainability/amazon-renewable-energy-goal 75 2023 Amazon Sustainability Report; https://sustainability.aboutamazon.com/2023-amazon-sustainability-report.pdf Attachment Staff Report Item 12C
13 of 18
35% of incremental power demand from Google and Microsoft.”76 As a demonstration of this surge in carbon-emitting sources, current wait times for new gas-fired turbines can be as long as 7 years, with costs likewise escalating, due to high demand from data centers.77 As the resource mix changes, hyperscalers are also shifting procurement strategies, moving from traditional PPAs to direct capacity investments, such as Google’s $4.75 billion acquisition of Intersect.
78 Activity amongst the largest hyperscalers suggest a willingness to spend liberally, so long as the needed energy and capacity are readily available. There currently exist a wide spectrum of procurement options for data centers, ranging from fully off the grid to cost-of-service ratemaking under an existing – or updated – industrial tariff. A list of these various options was produced as part of a December 2025 report by E3 and is reproduced as Table 1 below.79
Table 1. Sampling of Rate and Contract Design Options, with comparison of values. Key: tilde = partial utility service, check= full utility service; green = high, yellow = medium, and red = low score of the listed value.80
Contracting
Options Definition
Utility
servic
e
Promote
s Data
center
Growth
Risk
Burden
on Utility
(U) or
Develope
r (D)?
Protects
Existing
Customers
Shared
Infrastructure
Needed
Potential
Ratepayer
Benefits
Relative
Ease of
Impleme
ntation
Behind-the-Meter
Load co-located with
generation,
islanded/off-grid
D
Wholesale
Customer
Transmission
-level
customer
with minimal
utility service
~
D
Physical PPA
Customer
contracts
with
developer for
physical
power, utility
provides
T&D
✓
D
Virtual PPA Customer
contracts for
✓ D
76 B. Brunettil, et al., “Hyperscaler procurement to shape US power investment,” S&P Global, December 19, 2025; https://www.spglobal.com/sustainable1/en/insights/special-editorial/hyperscaler-procurement-to-shape-us-power-investment 77 Jared Anderson, “US gas-fired turbine wait times as much as seven years; costs up sharply;” S&P Global, May 20, 2025. https://www.spglobal.com/energy/en/news-research/latest-news/electric-power/052025-us-gas-fired-turbine-wait-times-as-much-as-seven-years-costs-up-sharply 78 Alphabet Investor Relations, “Alphabet Announces Agreement to Acquire Intersect to Advance U.S. Energy Innovation,” December 22, 2025; https://abc.xyz/investor/news/news-details/2025/Alphabet-Announces-Agreement-to-Acquire-Intersect-to-Advance-U-S--Energy-Innovation-2025-DVIuVDM9wW/default.aspx 79 E3, Tailored for Scale: Designing Electric Rates and Tariffs for Large Loads, December 2025; https://www.ethree.com/wp-content/uploads/2025/12/RatepayerStudy.pdf 80 Figures 7 and 8 (pg 17-18) in E3 Tailored, Ibid.

Attachment Staff Report Item 12C
14 of 18
renewable
energy
credits with
no physical
delivery of
power
Semi-Islanded/Non
-firm
Interruptible
service,
supplemente
d with on-site
generation
~
D
Rolled-in
Incremental
New
infrastructure
costs are
rolled into
existing rate
structure and
thus spread
across all
customers.
✓
U
As mentioned above, data center developers have been trying out these various procurement arrangements throughout the U.S. In California, activity has largely centered on individual project-by-project review for energization cost allocation, and traditional C&I tariffs for rate design. Other options listed in Table 1 are limited in California, such as the physical PPA design, which is subject to a statutory kWh limit and rarely open to new entrants.81 While data centers currently favor energization speed and build out, eventually the industry is likely to prioritize cost-effective energy arrangements to ensure continued growth; in which case alternate rate and contract designs will likely be explored more. For instance, under the right market conditions, one can envision data centers under existing IOUs energization agreements, exiting those agreements and turning to large amounts of on-site generation. The consequences of having large, heavily resourced firms seeking such alternate arrangements could be significant.

When the Backup Becomes the Main Act. It is also unclear, as hyperscalers are developing more fossil resources nationwide, whether that trend will also emerge in California. The CEC has exclusive authority to certify thermal power plants of 50 MW or more in California. The last natural gas plant the CEC permitted was in 2017;
82 however, they have consistently authorized
(via an exemption) diesel backup generating facilities co-located with data centers. Under the Small Power Plant Exemption (SPPE) process, applicants proposing facilities between 50 and 100 MW may seek an exemption from CEC certification. The CEC may grant an SPPE if it determines that the project would not result in significant adverse impacts to the environment or energy resources. If an exemption is granted, local land use authorities and other permitting agencies, such as the local air district, assume jurisdiction and act as responsible agencies under the California Environmental Quality Act (CEQA), conducting any additional environmental review required for their approvals.

Since the establishment of the SPPE process in the 1980s, the CEC has provided a total of 36 SPPEs, including 17 for data centers’ backup power generation. All the SPPE applications filed since 2011 have been for backup generating facilities serving data centers, and all but one of the 81 Public Utilities Code §365.1
82 Huntington Beach Energy Project; https://www.energy.ca.gov/powerplant/combined-cycle/huntington-beach-energy-project Attachment Staff Report Item 12C
15 of 18
17 data center exemptions were provided for diesel generator systems.83 Fuel costs, fuel availability, and energy capacity shape the selection of diesel generators over lower and zero-emitting backup power systems.

The CEC’s power plant siting authority was designed to provide a transparent and predictable certification process, including meaningful opportunities for public participation and consideration of alternatives under CEQA. By contrast, projects approved through the SPPE offer limited formal opportunities for stakeholder engagement, and the program was not designed 40 years ago with today’s large, data center–driven power systems in mind. Moreover, data centers in the state have concentrated within small, densely urban geographic areas, primarily in the Silicon Valley; as a result, these SPPEs may disproportionately impact certain communities near large population centers, increasing local air pollution concerns. Given the scale and pace of data center expansion, the joint committee may wish to consider whether the SPPE remains the right pathway for authorizing these facilities.

Moreover, as data center developers begin to embrace flexible demand in exchange for faster connectivity84– such as what is proposed in the recent FERC Advance Notice of Proposed Rulemaking (ANOPR)85– the potential for increased diesel usage could rise, adversely affecting local communities. Data centers have traditionally been considered inflexible loads, operating 24/7 to provide constant service such as cloud storage. However, certain AI workloads, such as AI training and machine learning are less time-sensitive than traditional data center workloads and can tolerate brief interruptions.86 The promise of flexible data center demand is that users capable of reducing their grid usage, especially during periods of peak energy usage, could significantly reduce resource and infrastructure needs. However, just because a data center goes off the electric grid doesn’t mean it goes dark. Depending on how rules for data center demand flexibility are written, data centers could still be considered flexible by switching their operations to their backup diesel units, greatly increasing their emissions. The joint committee may wish to ask panelists what protections could be developed to ensure encouraging flexible demand at data centers does not inadvertently lead to overuse of onsite diesel backup generators.

VI. Cost.
“Every houseguest brings you happiness – some when they arrive, and some when they are leaving.”87
Imagine roommates sharing an apartment who split rent and utilities evenly. When a new roommate moves in, if the new roommate pays at least their share of the added costs – utilities, wear and tear, groceries – the original roommates aren’t paying extra. If the new roommate pays less, the existing roommates subsidize them. If the new roommate covers not only their own costs but also contributes to fixed expenses like rent, everyone’s share goes down. Fairness 83 https://www.energy.ca.gov/programs-and-topics/topics/power-plants/power-plant-compliance-and-siting 84 Norris, T. H., T. Profeta, D. Patino-Echeverri, and A. Cowie-Haskell. 2025. Rethinking Load Growth: Assessing the Potential for Integration of Large Flexible Loads in US Power Systems. NI R 25-01. Durham, NC: Nicholas Institute for Energy, Environment & Sustainability, Duke University. https://nicholasinstitute.duke.edu/sites/default/files/publications/rethinking-load-growth.pdf 85 Chris Wright, Letter to FERC to Initiate Rulemaking “…Regarding the Interconnection of Large Loads Pursuant to the Secretary’s Authority Under Section 403…”, October 23, 2025. https://www.energy.gov/sites/default/files/2025- 10/403%20Large%20Loads%20Letter.pdf
86 Y. Numata, et al., “Fast, Flexible Solutions for Data Centers,” RMI, July 17, 2025; https://rmi.org/fast-flexible-solutions-for-data-centers/ 87 Aphorism widely attributed to Confucius
Attachment Staff Report Item 12C
16 of 18
suggests a new roommate should at least cover what they add and ideally contribute to existing costs.

The emergence of extensive data center electricity load, driven largely by the growth in artificial intelligence computing, marks a rare opportunity in the energy industry to reassess this contract amongst roommates. A large, well-financed new roommate wants to move in. Ensuring new large loads provide sufficient revenue to meet or exceed the marginal cost of service88 should protect existing utility customers.

It is the responsibility of the utility to provide electrical service to customers within their territory; known as their “obligation to serve.” Traditionally, the costs of providing that service, including potential grid upgrades, are socialized among all customers. However, existing rules for customers seeking to energize at the distribution-level89 include provisions that reduce risks of stranded costs for ratepayers and require some customers to pay for part of their own energization.90
Electric rate design generally rests on the principle that other ratepayers should not pay for upgrades from which they are not beneficiaries. Just how much an existing utility customer may or may not benefit from emergent data center energization is a topic of ongoing debate. Cost to energize new data centers are usually broken down by 1) the cost to connect them to the grid (their “energization” cost); and 2) the cost of continuing to provide them service (their rate schedule). For California’s IOUs, the traditional arrangement involves individual project-by-project review of the energization cost,91 and traditional C&I rate tariffs.

Electric Rule 30. In November 21, 2024, PG&E filed an application at the CPUC to establish a streamlined approach for energizing new transmission-level retail customers.92 While the application is broad in its definition of “large load customers” that would apply for service under this tariff, the PAO notes 75% of PG&E customer requests to interconnect at the transmission-level in the last two years are from data centers. That represents about 30 data center requests in 2 years, compared to 24 total customer requests for transmission-level service in the prior 8 years (2014-2022).93 The Electric Rule 30 proposal is solely focused on the energization costs, not a unique large load rate.

Under the proposed Rule 30:

1. Large load customers are required to pay upfront the cost of new interconnection facilities, or the customer can build the facilities herself and transfer ownership to PG&E, if more cost-effective. Costs for upgrades on the existing transmission system would be paid by ratepayers. PG&E states this is due to these facility upgrades benefitting multiple customers as well as advancing California policy initiatives such as electrification. 88 “Marginal cost of service” generally means the cost to the utility to serve an additional unit of electricity demand at a given time. These are the costs in power markets or the costs to build new infrastructure in serving new load. 89 CPUC Electric Rules 15 and 16
90 Essentially, customers are given a set “allowance” which is paid for by all ratepayers. Any costs in excess of the allowance are covered by the individual seeking to energize. The allowances are formula amounts set by the CPUC. 91 Via “Exceptional Case Filings” such as the recent approval of a 90MW Microsoft data center. https://docs.cpuc.ca.gov/PublishedDocs/Published/G000/M590/K889/590889612.PDF 92 A. 24-11-007
93 Pg. 2, “PAO Opening Brief,” A. 24-11-007, filed October 24, 2025.

Attachment Staff Report Item 12C
17 of 18

2. Large load customers are required to pay the actual costs of interconnection infrastructure, rather than the estimated costs, which are traditionally used for residential and commercial customer energization.

3. After the large load customer has received service and is providing revenues, the customer is eligible for a refund of their upfront payments over time. Refunds are based on the revenues produced by the large load customer. Thus, if the customer’s load does not materialize, the customer does not receive back its entire upfront payment. A large load customer only receives refunds when it is producing revenues to offset the interconnection infrastructure costs. However, parties have raised that the timing of that repayment should be matched with the timing of the new revenue generated from the customer, such that annual refunds never exceed the net customer revenue.

4. Includes a minimum demand charge, minimum contract term, and early termination fee.

5. Information sharing provisions to facilitate community choice aggregators (CCAs) and PG&E having early notice of potential large load customers and being able to make cost-effective generation resource decisions.

While PG&E notes in its application that these provisions will protect existing ratepayers and ensure the revenues from large load customers result in bill reductions, others within the proceeding disagreed and proposed a number of modifications. 94 In July 2025, the CPUC
adopted interim implementation of PG&E’s Rule 30, requiring new transmission-level customers to be responsible for all initial costs, deferring decisions on rate recovery and cost allocation until a final decision, and denying PG&E’s request for a memorandum account to record accrued interest, among other requirements.
95 Do Large Loads Mean Lower Rates? The promise of data center energy consumption is alluring: that increased electricity sales from data centers will cover all the new costs to serve those data centers and may even offset existing system costs. Such is the story for other large loads, like transportation electrification (TE), with the California Public Advocates Office (PAO) finding TE “may cause downward pressure on electric rates.”96 The timing can feel almost too perfect: as California undertakes systemwide decarbonization and infrastructure renewal, a new, well-funded customer emerges, apparently willing97 to shoulder the cost of long-deferred investments. However, as the PAO cautions for TE, and as applies equally to data-center development, “to achieve this downward pressure on rates, effective management of multiple factors will be required, including efficient infrastructure buildout and cost constraints.”98 Rate design is an essential mitigation tool to ensure costs are not shifted onto other ratepayers.

The treatment of large loads under retail ratemaking presents a challenge, as the pace of retail design has traditionally been aligned with smaller, incremental load growth. 94 See “PAO Opening Brief,” A. 24-11-007, filed October 24, 2025. 95 D. 25-07-039. CPUC “Decision Partly Granting and Partly Denying Pacific Gas and Electric Company’s Motion for Interim Implementation of Electric Rule Number 30,” A. 24-11-007, July 24, 2025; https://docs.cpuc.ca.gov/PublishedDocs/Published/G000/M574/K875/574875643.PDF 96 Pg. 14, Public Advocates Office, Distribution Grid Electrification Model 2025 – Study and Report, October 2025. https://www.publicadvocates.cpuc.ca.gov/-/media/cal-advocates-website/files/press-room/reports-and-analyses/251030-public-advocates-office-distribution-grid-electrification-model-2025.pdf 97 Amrith Ramkumar, “Microsoft Makes New Data-Center Pledges After Local Backlash,” The Wall Street Journal, January 13, 2025.
98 Pg. 14, PAO October 2025, Ibid.

Attachment Staff Report Item 12C
18 of 18
PG&E has provided “extensive analysis” in its Electric Rule 30 application showing that the substantial revenues from large load customers will result in bill reductions for existing ratepayers because fixed costs will be spread over a larger base of customers.99 While the simplicity of the math is undeniable – larger denominator, smaller overall number – it is unclear if their analysis truly considers all known costs. PG&E’s analyses do not appear to be part of the public docket by which the committee might review.

Transmission-level customers, such as large data centers, typically pay rates about 14-17¢ less than distribution-level customers.
100 Distribution rates also contain statewide policy costs, such as wildfire mitigation, net energy metering, and public purpose program costs. Customers connecting at the transmission-level do pay certain distribution costs, including a portion of wildfire-related costs,101 but not nearly as significantly as distribution-level customers.

The primary way large loads reduce bills for other ratepayers is by (a) minimizing or avoiding infrastructure upgrades that would otherwise be socialized – either by limiting the need for new infrastructure or requiring the customer to pay those costs upfront – and/or (b) ensuring that large-load customers provide stable, long-term revenue over the life of any required upgrades, sufficient to cover not only fixed capital costs but also ongoing operations, maintenance, depreciation, and associated generation procurement costs.

The joint committee may wish to ask panelists how to structure tariffs to balance encouraging economic development while guarding against higher energy costs to existing ratepayers, stranded assets, and negative system impacts.

Conclusions. The ongoing surge of investment into AI data centers has been suggested to be bigger than the buildout of the interstate highway or the dot-com boom. AI-driven data center growth presents California with both a stress and a strategic choice. The scale, speed, and uncertainty of this new load are unlike prior waves of electrification, with the potential to challenge long-standing assumptions embedded in forecasting, planning, cost allocation, and rate design. California’s relative insulation to date reflects deliberate policy choices, higher baseline prices, and robust planning institutions; but those advantages are not guaranteed if safeguards do not evolve as rapidly as the load itself. The core question is not whether data centers should be served, but under what terms: who pays, when, and with what risks. If large loads are required to bear the full costs they impose, provide durable and verifiable revenue, and align procurement and operational practices with reliability and climate goals, they could support system investment without burdening existing customers. Absent those protections, the promise of “downward pressure on rates” risks becoming another case of privatized benefits and socialized costs. The task before policymakers is to ensure that California welcomes innovation without repeating the mistakes now playing out elsewhere; and that, when a powerful new roommate moves in, the household is better off, not worse.
# # # # # # #
99 Michael Medieros, “PG&E Written Testimony;” Letter to the Little Hoover Commission; December 8, 2025. https://lhc.ca.gov/wp-content/uploads/PGE-LHC-Written-Testimony.pdf 100 Pg. 26, “Little Hoover Commission Study to Review Data Centers and California Energy Policy, Pt. II,” December 11, 2025. https://lhc.ca.gov/wp-content/uploads/Revised-Commissioner-Information-12-11-25.pdf 101 Medieros, Ibid.

Attachment Staff Report Item 12C
Staff Report Item 13
To: Ava Community Energy Authority
From: JP Ross, Vice President Local Development
Subject: Staff Report updating the Board on the launch of Ava’s SmartHome Battery Program
Date: February 18, 2026
__________________________________________________________________ Summary/Recommendation
This Staff Report updates the Board on the launch of Ava’s SmartHome Battery Program Financial Impact
N/A
Analysis and Context
Ava has allocated $15M to the SmartHome Battery and Resilience Hub initiative. The Board approved the Program Design and Budget at the November 2025 Board meeting. Staff will provide an update on the launch timing and activities of the SmartHome Battery program.

Attachments
A. Presentation: SmartHome Battery Pre-Launch Update
SmartHome Battery
& Resilience
Hubs Updates
February 2026
JP Ross, VP Local Development
Attachment Staff Report Item 13A
Agenda
2 • Program Overview
• Launch Timing
• Installer and Customer Engagement
• Ava’s Virtual Power Plant (VPP)
• Resilience Hubs
Attachment Staff Report Item 13A
Program Goal
● Deploy $15M in upfront and ongoing incentives that
support community resilience, adding up to 21MW of
dependable powered batteries to our community Virtual
Power Plant (VPP).

Program Audiences
● Residential Customers – SmartHome Battery
 Income qualified (CARE/FERA)
 General Market
● Resilience Hubs
Program Learnings
● Enhance our understanding of how to reliably manage
and grow battery aggregations
● Understand what is needed to deploy Community
Resilience Hubs
Program Goals, Audience and Learning
3 Attachment Staff Report Item 13A
SmartHomeBattery + Resilience Hubs Launch Timing
4 SmartHome Battery (Residential Customers)
• March 2026: Program opens for newly installed solar
and battery systems
• April 2026: Program opens for customers
with existing solar and battery storage systems
• Q2/3 2026: Program opens for customers whose
batteries are owned by aggregators pending
aggregator execution of a participation agreement
with Ava.

Resilience Hubs
• January 2026 onward: Resilience hub recruitment for
Technical Assistance
• Q2: Technical Assistance available for Resilience
Hubs
• 2H 2026: Targeting Resilience Hub Incentive launch
2025
Resilience Hub
Education and
Awareness Grant
March ‘26
Residential Launch
Q2 26
Resilience Hub
Technical
Assistance Launch
SHB Aggregator
launch*
2H 2026
Resilience Hub
Incentive Launch
Attachment Staff Report Item 13A
Ava is engaging with installers to both raise
awareness and support installers to learn how to
enroll customers in SmartHome Battery.

Activities include:
• Virtual Installer Informational Webinar (January
30)
• In-person Installer Training 1 in Oakland (February
25)
• Two in person additional installer trainings
(Forthcoming in 2026, across service area)
• Engagement with California Solar and Storage
Association (CALSSA)
SmartHomeBatteryInstaller and Customer Engagement
5 Installer Engagement Customer Engagement
Ava will announce SmartHome Battery launch
across our owned channels and across key
stakeholders. Installers will be the main channel.

Activities include:
• Posts on Instagram, Facebook, Bluesky, and
LinkedIn
• Inclusion in Direct Current Newsletter and CEO's
Desk Newsletter
• Social/email content to share with Munipals
• Inclusion in the community -based organization
newsletter
• Press Release
Attachment Staff Report Item 13A
Solar and Storage as a Part of Ava’s VPP
Ava’s DERMS allows us to leverage distributed energy resources (DERs) to reduce our carbon footprint and support customer savings from DERs by aggregating these technologies and optimizing them to operate in concert. This is known as a “virtual power plant” (VPP).

Residential solar/battery projects will be part of Ava’s Virtual Power Plant (VPP). 6 *
*Eight CMF projects not
eligible for incentives
Attachment Staff Report Item 13A
Ava's Virtual Power Plant
7 • Ava is working to add as many options as
possible for customers to choose
equipment.
• Initially Resilience Hubs will not be
required to enroll in Ava's VPP based on
lack of participation from commercial
battery and inverter OEMs
• The SmartHome Battery program will
launch with Tesla and Lunar batteries.

Agreements with several other OEMs are
underway and expected to be available
within a quarter of program launch.
• Integrating batteries with a DERMs
requires work from OEMs, can take up to
2 months, and requires extensive testing.
• Once operational maintaining battery
connectivity requires constant reporting.
*Ava's VPP and Resilience Hubs will not be Community Micro -Grids Attachment Staff Report Item 13A
Community Resilience Hub Initiative Review
8 Objective: Unde rstand the unique needs of community sites and provide them with resources, technical assistance, and incentives to develop local Resilience Hubs.

Community Outreach and
Site Identification ($300k)
Underway:
✓ Case Studies
✓ Resilience Readiness Tool
✓ Held 3 workshops with >150
registered participants and 26
CBOs to gather feedback
q Will lead 3 Community outreach to
identify interested sites
✓ One session complete 1/26
Technical
Assistance ($2M)
Forthcoming Q2:
• Resilience project sizing and
feasibility assessment
• Contactor bid reviews
• Available grants and incentives
• Additional services as needed
include: site visits, technical
contract agreement review, EV
charging assessment
Incentives
($3.75M)
Forthcoming 2H 2026:
• Upfront incentive
• Waitlist for projects as they
install equipment
• Incentive payment
• Potential for Ongoing VPP
incentives with battery OEM
integrations
Attachment Staff Report Item 13A
Thank you!
AvaEnergy.org
+1 833.699.3223
customer [email protected]
PoweredWithAva
Online
Phone
Email
Social
9 Attachment Staff Report Item 13A

Ava Community Energy Board of DirectorsFebruary 18, 2026

Ava Community Energy Board of Directors
February 18, 2026